University of Warwick institutional repository: http://go.warwick.ac.uk/wrap
A Thesis Submitted for the Degree of PhD at the University of Warwick
http://go.warwick.ac.uk/wrap/55431
This thesis is made available online and is protected by original copyright. Please scroll down to view the document itself. Please refer to the repository record for this item for information to help you to cite it. Our policy information is available from the repository home page.
A Multidimensional Analysis of
Post-Acquisition Performance:
The Case of Research and Development
in the Pharmaceutical Sector
Rupert Julian Booth
Thesis submitted for the degree of Doctor of Philosophy
in Business & Management Studies
University of Warwick, Coventry
Warwick Business School
September 2011 Contents
Acknowledgments xii
Declaration xiii
Abstract xiv
Glossary xv
1 Introduction 1
1.1 Research Aim 1
1.2 Choice of Sector and Need for Research 3
1.3 Research Approach 4
1.4 Research Hypotheses 5
1.5 Technical Challenges and Contributions to Knowledge 7
1.5.1 Measurement of Intangibles 7 1.5.2 Longitudinal Nature of R&D Pipeline 7 1.5.3 Variety of Outputs 8 1.5.4 Types of Contributions to Knowledge 8
2 Overview of Pharmaceutical Sector 10
2.1 Introduction 10
2.2 R&D Processes 10
2.3 Market Structure and Acquisition Activity 13
2.4 Key Metrics 14
ii 3 Literature Review 16
3.1 Introduction 16
3.2 PAP 16
3.2.1 History 16 3.2.2 Meta-Analyses 17 3.2.3 Summary of Main Approaches 24 3.2.4 Motive and Synergy 26 3.2.5 Diversification Literature 27
3.3 RBV 29
3.3.1 Early Definition of the RBV 29 3.3.2 Qualification of Resources 30 3.3.3 Dynamic RBV 31 3.3.4 Critiques of the RBV 31 3.3.5 Recent Retrospective on the RBV 33 3.3.6 Summary of Key Issues for Performance Measurement 34
3.4 PMFs 35
3.4.1 Theoretical Benefit of Additional Information 35 3.4.2 Benefits of PMFs 36 3.4.3 The Balanced Scorecard and its Evolution 37 3.4.4 Choice of Measures 40 3.4.5 External Evaluation of Intellectual Capital 41
3.5 Use of the RBV to Measure Performance in
Pharmaceuticals 44
3.6 Relevant DEA Literature 48
3.7 Synthesis 50
4 Design of PMF 52
4.1 Introduction 52
iii 4.2 PMF Design Principles 52
4.3 Availability of Information 56
4.3.1 Practicality 56 4.3.2 Scope and Structure 56 4.3.3 Resources and Barriers 57 4.3.4 Processes and Positions 59 4.3.5 Efficiency 60
4.4 Application of Design Principles to the Pharmaceutical
Sector 61
4.4.1 Demonstration of Use 61 4.4.2 Scope and Structure 61 4.4.3 Resources and Barriers 63 4.4.4 Processes and Positions 64 4.4.5 Efficiency 65
4.5 Proposed Pharmaceutical PMF 66
4.5.1 Outcome of Design Principles 66 4.5.2 Use of Negative Indicators 68 4.5.3 Choice of Financial Measures 68
4.6 Focus on R&D and Adaptation to DEA 69
4.7 Conclusion 71
5 Research Methodology 72
5.1 Introduction 72
5.2 Efficiency and Productivity Analysis Approaches 72
5.3 Summary of DEA Model 74
5.3.1 Orientation of Model 74 5.3.2 Inputs and Outputs 75
iv 5.4 Longitudinal Dimension 76
5.5 Definition of Notation 77
5.5.1 MCT 77 5.5.2 Normalisation Factor 78 5.5.3 Algebraic Notation for Efficiency Analysis 79 5.5.4 Form of DEA Equations 81 5.5.5 Algebraic Notation for Examination of Returns to Scale 82 5.5.6 Algebraic Notation for Classification of Acquisition History 83 5.5.7 Algebraic Notation for Statistical Testing 84
5.6 DEA 85
5.7 CCR Model 86
5.8 BCC Model 87
5.9 Output Weight Restrictions 88
5.10 Input Weight Restrictions 89
5.11 Returns to Scale 89
5.12 Design and Population of Acquisition
Typology 90
5.12.1 Identification of Acquisitions 90 5.12.2 Classification of Acquisitions 92 5.12.3 Identification of Cross-border and Cross-sector Acquisitions 92 5.12.4 Linkage to Major Pharmaceutical Companies 94
5.13 Analysis of M&A History 95
5.14 Statistical Test Approach 96
5.15 Incomplete Data 99
6 Data and Descriptive Statistics 100
6.1 Introduction 100
v 6.2 Available Input Data for the DEA Models 101
6.3 Available Output Data for the DEA Models 101
6.4 Comparing CRS and VRS Efficiency Scores 102
6.5 DEA Model Results: Comparing Input Assumptions 102
6.6 Financial Efficiency Data 103
6.7 Acquisition and NDV Data 103
6.8 Association of M&A and Technical Efficiency 103
6.9 Association of M&A and Financial Efficiency 104
6.10 Descriptive Statistics of the DEA Model Outputs 104
6.11 Descriptive Statistics of the R&D Process 106
6.12 Descriptive Statistics of M&A 109
6.13 Summary 112
7HypothesisTesting 113
7.1 Scope of Hypothesis Testing 113
7.1.1 Classical Hypothesis Testing Model 113 7.1.2 Set 1, Hypothesis 1: Returns to Scale 115 7.1.3 Set 2, Hypotheses 2, 3 & 4: Firm Acquisition History and Technical Efficiency 116 7.1.4 Set 3, Hypotheses 5, 6 & 7: Deal History and Financial Efficiency 117 7.1.5 Set 4, Hypotheses 8, 9 & 10: Acquisition History and Sectoral Efficiency 118 7.1.6 Set 5, Hypotheses 11, 12, 13 & 14: Acquisition History and Sales over Assets 120
7.2 Returns to Scale (H1) 121
7.3 M&A and Technical Efficiency (H2, H3, H4) 122
7.3.1 Hypothesis 2 122 7.3.2 Hypothesis 3 123
vi 7.3.3 Hypothesis 4 124 7.3.4 Summary of the Set 125
7.4 M&A and Financial Efficiency (H5, H6, H7) 126
7.4.1 Hypothesis 5 126 7.4.2 Hypothesis 6 127 7.4.3 Hypothesis 7 128 7.4.4 Summary of the Set 129
7.5 Sector Effects and Technical Efficiency (H8, H9, H10) 130
7.5.1 Hypothesis 8 130 7.5.2 Hypothesis 9 130 7.5.3 Hypothesis 10 131 7.5.4 Summary of the Set 132
7.6 M&A and Financial Metrics (H11, H12, H13, H14) 132
7.6.1 Hypothesis 11 132 7.6.2 Hypothesis 12 133 7.6.3 Hypotheses 13 & 14 133 7.6.4 Summary of the Set 135
7.7 Summary of Findings 135
7.7.1 Collation of Findings 135 7.7.2 Check for Consistency 136
8 Discussion of Results 139
8.1 Range of Findings 139
8.2 Returns to Scale of R&D 140
8.3 The Association of Technical Efficiency with M&A
History 141
vii 8.4 The Association of Financial Efficiency with M&A
History 142
8.5 The Sectoral Consequences of M&A 143
8.6 Relationship between SOA and M&A 144
8.7 Synthesis of Findings 145
9 Contributions 149
9.1 Overview 149
9.2 Contributions to the Acquisition Literature 150
9.2.1 Reduction of Uncertainty 151 9.2.2 Differences in Measured Performance 151 9.2.3 Diversification 153
9.3 Contributions to the Performance Literature 153
9.4 Contributions to the DEA Literature 154
9.5 Empirical Contributions 154
9.5.1 Returns to Scale 154 9.5.2 Power Laws in M&A 155 9.5.3 Financial Metrics in M&A 156
9.6 Directions of Future Research 157
10 References 158
viii A. R&D Data 182
A.1 Introduction 182
A.2 Historical R&D Data 182
A.3 Historical Sales Data 185
A.4 Current R&D & Composite R&D Data 188
A.5 Conclusion 193
B. Staff Data 194
B.1 Introduction 194
B.2 Staff Data 194
C. Power Laws 198
C.1 Introduction 198
C.2 Definition of a Power Law 198
C.3 Alternative Distributions 199
C.4 Relevance to this Thesis 200
D. DEA Model Data 201
E. AcquisitionData 242
E.1 Introduction 242
E.2 Search Criteria 242
E.3 Annual Analysis 244
E.4 Detailed Merger Data 246
ix List of Tables
3.1 Choice of Measures in Acquisition Meta-Analysis 1983–2006 ...... 19 3.2 Contributions of the Major RBV Authors to Performance Measurement.... 34 3.3 Summary of Balanced Scorecard Concepts ...... 39 3.4 Parameters and Metrics Cited in DeCarolis & Deeds (1999) ...... 45 3.5 Outcomes of Hypotheses Testing in DeCarolis & Deeds (1999) ...... 46 3.6 Parameters and Metrics in DeCarolis (2003) ...... 47 3.7 Outcomes of Hypotheses from DeCarolis (2003) ...... 48
4.1 Design Principles ...... 53 4.2 Proposed Pharmaceutical PMF ...... 66 4.3 Scorecard for Pharmaceutical Company ...... 67
5.1 Time and Costs of R&D Development ...... 76 5.2 Algebraic Notation for DEA Models ...... 80 5.3 Algebraic Notation for Examining Returns to Scale ...... 82 5.4 Algebraic Notation for Classification of Acquisition History ...... 83 5.5 Algebraic Notation for Hypothesis Testing ...... 85 5.6 Selection Criteria ...... 91 5.7 Key Terms for Acquisition Typology ...... 96
6.1 Descriptive Statistics of Ratio of Trials to Compounds ...... 105 6.2 t test on Ratios of Trials to Compounds in Successive Phases ...... 105 6.3 Summary of Association of Acquisition History with Technical Efficiency ...... 112 6.4 Summary of Association of Acquisition History with Financial Efficiency ...... 112
7.1 Results of Tests on Hypothesis H1a ...... 121 7.2 Results of Tests on Hypothesis H2a ...... 122 7.3 Results of Tests on Hypothesis H3a ...... 124 7.4 Results of Tests on Hypothesis H4a ...... 125 7.5 Results of Tests on Hypothesis H5a ...... 127 7.6 Results of Tests on Hypothesis H6a ...... 128 7.7 Results of Tests on Hypothesis H7a ...... 129 7.8 Results of Tests on Hypothesis H8a ...... 130 7.9 Results of Tests on Hypothesis H9a ...... 131
x 7.10 Results of Tests on Hypothesis H10a ...... 131 7.11Results of Tests on Hypothesis H11a ...... 132 7.12 Results of Tests on Hypothesis H12a ...... 133 7.13 Results of Tests on Hypothesis H13a ...... 134 7.14 Results of Tests on Hypothesis H14a ...... 134 7.15 Rejected Null Hypotheses...... 135 7.16 Boderline Alternative Hypotheses ...... 137
A.1 Historical R&D Data in $million Actual ...... 182 A.2 Historical Revenue Data in $million Actual ...... 185 A.3 Current & Composite R&D Data in $million Actual...... 190
B.1 Staff Numbers ...... 194
D.1 DEA Model Inputs ...... 201 D.2 DEA Model Compounds (Comp.) Output Data ...... 205 D.3 DEA Model Clinical Trials Output Data ...... 208 D.4 DEA Efficiency Scores for Compounds as Outputs ...... 211 D.5 DEA Efficiency Scores for Trials as Outputs ...... 214 D.6 DEA Efficiency Scores for VRS for Alternative Input Assumptions ...... 217 D.7 ROS, ROA and SOA...... 220 D.8 SDV, Cost of Sales and NDV Data for Firms ...... 223 D.9 NDV and Pure Technical Efficiency (Base Model) ...... 226 D.10 NDV and Pure Technical Efficiency (Staff Input) ...... 228 D.11 SDV and Pure Technical Efficiency (Base Model) ...... 230 D.12 Association of NDV and ROS ...... 233 D.13 Association of NDV and ROA ...... 236 D.14 Association of Acquisition History and SOA ...... 239
E.1 Database Search Criteria ...... 243 E.2 Date Analysis ...... 245 E.3 Detailed Merger Data ...... 247
List of Figures
4.1 Summary of Design Principles...... 55 6.1 Graph of Scale Efficiency Versus Mean R&D (Compounds as Output) .. 107 6.2 Graph of Scale Efficiency Versus Mean R&D (Trials as Output) ...... 107 6.3 Distribution of R&D Expenditure of Above- and Below-Median Efficiency 108 6.4 Frequency of M&A Deals by Value ...... 109 6.5 Normality Test for Square Root of NDV less Zeros ...... 111
xi Acknowledgements
I would like to thank my supervisors Dr Duncan Angwin and Dr Mette Asmild for their assistance and encouragement and to acknowledge the example of my late mother Dr Pauline Rita Booth (1930–2009).
xii Declaration
I declare that the content of this thesis is entirely my own work and has not been submitted for a degree at another university.
xiii Abstract
This thesis provides an additional perspective of the Merger Paradox, namely that mergers and acquisitions (M&A) continue to be transacted when historically their results seem to be disappointing overall.
The thesis shows that when a theoretically sound basis (related to the Resource Based View and expressed as twelve design principles) is used to design a performance measurement framework, then there is no association between a firm's post-acquisition performance and the scale of a firm's previous acquisitions; the thesis then shows, by contrast, that there is a positive association between firms with an above-average level of past acquisitions (by value) and higher financial performance. This divergence provides both a motive and an ability to continue to undertake M&A, despite a lack of association of acquisitions with longer-term operational performance and very strong evidence of diseconomy of scale in the most crucial business process, for the case examined, which is the research and development (R&D) process in the research-based pharmaceutical sector. Additionally, the thesis examines the relative merits of Return on Sales and Return on Assets as financial metrics of performance, and establishes statistically significant differences in the measurement of performance by these two metrics.
The thesis also establishes a contrast between the findings at the level of the firm and at the level of the sector, namely acquisitions considered in aggregate are associated with gains at the sector level, even though this association was not observed when acquisition was considered at the level of the acquiring firm.
The thesis provides a new application of Data Envelopment Analysis and establishes a scale efficiency relationship for the pharmaceutical R&D process. A further empirical contribution is the examination of the statistical distribution of acquisitions in the pharmaceutical sector and confirmation of the consistency of that distribution with a power-law.
xiv Glossary
ADME Absorption, Distribution, Metabolism and Excretion
BCC Banker, Charnes and Cooper
CCR Charnes, Cooper and Rhodes
CRS Constant Returns to Scale
DEA Data Envelopment Analysis
DMU Decision Making Unit
DRS Decreasing Returns to Scale
IC Intellectual Capital
ICT Information and Communications Technology
IND Investigational New Drug
IRS Increasing Returns to Scale
KPI Key Performance Indicator
M&A Mergers and Acquisitions
MCT Measure of Central Tendency
Merger Paradox M&As continue to be transacted when historically their
results seem to be disappointing overall
NAIC North American Industry Code
NDA New Drug Application
xv NDV Normalised Deal Value p-value The probability, computed assuming that the null
hypothesis is true, of observing a value of a test statistic
that is at least as contradictory to the null hypothesis as
the value actually computed from the sample data
(adapted from Bowerman et al., 2011: 359)
PAP Post-Acquisition Performance
PMF Performance Measurement Framework
R&D Research & Development
R&D (Average) Average Research & Development expenditure for a firm
from 2001 to 2006
R&D (Current) Research & Development expenditure for a firm in 2006
R&D (Historic) Average Research & Development expenditure for a firm
from 2001 to 2005
RBV Resource Based View
ROA Return on Assets
ROE Return on Equity
ROS Return on Sales
SDV Sum of Deal Value
SOA Sales over Assets
xvi US$ United States dollar. Official currency of the United
States of America and several other countries
VRIN Valuable to the company, Rare, Imperfectly imitable and
Non-substitutable
VRS Variable Returns to Scale
xvii 1 Introduction
1.1 Research Aim
Angwin (2007) posits a fundamental question for research into mergers and acquisitions (M&A), which is often termed the ‘Merger Paradox’, namely why
M&A continue to be transacted when historically their results seem to be disappointing overall. That paper goes on to suggest that there are many objectives for M&A and one should not judge a transaction to be a failure on the basis of a particular measure of performance. This thesis examines the
Merger Paradox from the stance of measurement rather than from the motive itself (although the two are related), and seeks to demonstrate that differences in the measures used to determine post-acquisition performance (PAP) could explain the continuing popularity of M&A despite M&A not being associated with improved long-term performance in crucial business processes.
The literature review for this thesis has identified PAP literature dating back to
1968 and from the very start the issue of multiple stakeholders and dubious
PAP was fully recognised. Over 40 years of research later, across several disciplines, Zollo & Meier (2008) noted the absence of a convergence of findings even within disciplines and identified the use of 12 types of measures of PAP. Despite the variety of measures, in a recent investigation using multiple criteria, Papadakis & Thanos (2010) noted disappointing outcomes in over half of cases, which then leads to the Merger Paradox of why acquisitions remain popular when, in most cases, they seem to be unsuccessful.
It is against this background of voluminous, diverse but pessimistic literature that this thesis adds new contributions by initially focusing on the principles of performance measurement, noting that PAP is an intellectual construct and
1 taking heed of Venkatraman & Grant (1986) who observe that in strategic management the principle of measurement of a construct is often ignored in favour of content development. In essence, PAP itself can never be considered directly, but only through proxies for PAP expressed in a chosen measure. Given this, the selection of any measure requires a theoretical basis.
The Resource Based View (RBV) of a firm is used here as the theoretical basis for measure selection. Currently, RBV is a long-established approach to strategic management. The Journal of Management in September 2011 dedicated a special issue to reviewing resource based theory to commemorate its previous special issue that introduced the theory 20 years earlier; in the most recent special issue, Barney et al. (2011) considered it had reached maturity and was capable of further development to satisfy its critics.
The focus of the RBV is on the linkage of competitive advantage to the differences between firms in the same market, as opposed to the factors affecting profitability in the market as a whole (the focus of the Industrial
Organisation approach to strategy). The focus of RBV on relating the competitive advantage of a firm to its special factors (some of which will be measurable) has made its literature a natural basis from which to develop a systematic means to identify a set of performance measures that can be related to long-term performance. In the thesis, a performance measure is used to assess the comparative efficiency of key processes of firms in a particular sector and it is used as a measure of relative non-financial performance. This performance measure is then associated with the acquisition history of the firms and used to test a series of hypotheses for mergers in aggregate as well as cross-sector and cross-border mergers. In order to shed new light on the Merger Paradox, the outcome of the analysis is compared with a similar exercise using a common accounting measure.
2 Finally, the effect of M&A on sector performance and whether particular alternative financial measures provide a consistent view on PAP are examined.
1.2 Choice of Sector and Need for Research
The research-based pharmaceutical sector offers several advantages for this research:
– The resources of a firm, in the RBV sense that includes both inputs and
outputs of the research and development (R&D) process, are clearly
defined because products cannot be developed without formal regulation
and identification, nor can they be sold without marketing approval.
– The R&D process is generic across the whole industry (because of
regulatory constraints) and this gives rise to a comparable process that
allows efficiency to be measured.
– Data on the industry are widely available.
These characteristics do not apply to all research-based industries, for example the Information and Communications Technology (ICT) sector has diverse R&D processes.
Furthermore, the research-based pharmaceutical sector offers compelling public policy reasons to undertake research: rising healthcare costs are a feature of the economies of the developed world and the ethical pharmaceutical sector is both a cause of, and a potential solution to, these costs; yet there is concern that it is facing a productivity crisis, for example
Cockburn (2006), and a rising cost per new compound, for example DiMasi et al. (2003).
3 1.3 Research Approach
Firstly, because the focus of the research is on how the assessment of PAP depends on the choice of measure, it is necessary to first develop a systematic and rigorous way to select performance measures; as indicated in
Section 1.1, this is based upon the RBV. The selection process leads to a vector of measures that can be used to measure R&D efficiency.
Secondly, having defined a vector of measures, a number of Data
Envelopment Analysis (DEA) models were built with the selected measures to measure comparative efficiency of the R&D process between firms, in order to establish relative performance; in order to do this it was necessary to test for returns to scale, after populating the models with data on the inputs and outputs to the process. The different models had different selections of inputs that gave rise to different assessments of efficiency.
Thirdly, data on acquisitions of the major pharmaceutical firms were collected and the deals were also classified according to whether the acquisition involved diversification into different nations or sectors. The data on M&A values were summated for each of the firms and also normalised by dividing by the cost of sales of the firm to relate the deals to the size of the firm.
Fourthly, this acquisition history of the firms was associated with their efficiency as measured by DEA and used to test three merger-related hypotheses relating to: acquisitions in total, acquisitions of new product resources, and acquisitions of new market resources. This was done for more than one DEA model and the differences in outcomes were related to typical behaviours following a merger.
4 Fifthly, the financial efficiency of the firms was measured by Return on Sales
(ROS) and Return on Assets (ROA), for the same firms and this was used to test the same three hypotheses. This was then compared with the outcomes of the hypothesis tests using the DEA methods and this led to a further explanation of the Merger Paradox.
Sixthly, the acquisition statistics on mergers of the firm were analysed without the application of a normalisation factor to examine the effect of mergers on efficiency at the sectoral level; this was done in order to compare the outcome with financial methods of analysis in which differences are observed for the
PAP of the shareholders acquiring the firm alone and the shareholders of the acquiring and acquired firm together.
Seventhly, the statistic Sales over Assets (SOA) was tested in order to understand if the acquisition process had an effect on common financial metrics that could be used to measure PAP and an effect was detected.
Some of these steps required hypothesis testing. The research hypotheses are described in the next section (1.4). The research hypotheses were developed after consideration of the literature on PAP and will be related to the literature in the final chapters of the thesis.
1.4 Research Hypotheses
The hypotheses are based on testing the difference in a measure of central tendency (MCT) between two samples, using both a parametric and a non- parametric test. All the hypotheses follow a conventional format, whereby the null hypothesis represents the case of no difference between the means of two samples; the alternative hypothesis is therefore that there is a difference between the means of the two samples and it is the alternative hypothesis that
5 is tested to establish a statistically significant difference between the means of the two samples.
In the case of the first set of hypotheses, there is a single null hypothesis that scale does not vary with size: constant returns to scale (CRS) exist for a process. The alternative hypothesis is that there are variable returns to scale
(VRS) and this is tested by examining whether statistically significant differences in size exist between two groups comprising firms with above- median and below-median efficiency scores.
The second set of hypotheses (a set of three) examines technical efficiency.
The null hypothesis is taken to represent the situation that a history of M&A transactions, normalised for the size of the firm, is not associated with a change in efficiency. The alternative hypothesis is that efficiency does change as a result of M&A and as the objective is to test the Merger Paradox, the statistical test is one-sided: M&A is associated with lower technical efficiency.
A similar test is undertaken for a third set of three hypotheses that examines financial efficiency (as measured by both ROS and ROA), except that the direction of the alternative hypothesis is that M&A is associated with higher
ROS and ROA that would provide a financial motive, or at least a qualifying factor, for the deal.
A fourth set of three hypotheses considers the association of acquisitions in total, without normalisation for the size of the firm, with technical efficiency.
These hypotheses follow a similar format to those in the second set.
Finally there is a fifth set hypotheses that examines different financial metrics.
The null hypothesis is that SOA is unchanged, and the alternative hypothesis is that M&A is associated with lower SOA, as additional intangible assets
6 become recognised in the M&A process. A fourth hypothesis is added to this set, a non-directional hypothesis, to clarify one of the findings.
There are therefore 14 hypotheses in all.
1.5 Technical Challenges and Contributions to Knowledge
1.5.1 Measurement of Intangibles
The difficulty in quantifying intangibles has posed various challenges to this research. One example of an intangible factor is the nature of the output of the
R&D process, however, the sector was chosen so a regulatory process circumvented the difficulty of recognising the worth of an output. A further aspect relates to the attempts to measure diversification, for example M&A deals have been categorised in order to examine the effects of diversification, however in order to do so there needs to be a measure of relatedness. A measure of relatedness was achieved by identifying whether the acquired company was located in the same country or had the same industrial classification as the acquiring company; nonetheless it is recognised that these simple classifications do not account for a more nuanced situation.
1.5.2 Longitudinal Nature of R&D Pipeline
The use of DEA to measure the comparative efficiency of the pharmaceutical
R&D pipeline between firms in the M&A context is novel1. One possible reason is because of the difficulties presented by the longitudinal nature of the pharmaceutical R&D pipeline, whereby outputs in the current time period depend on inputs in the previous time periods. This research has sought to address this problem by collecting input data that cover the majority of the
1 It has however been used as a means of R&D productivity measurement in other sectors, unrelated to
M&A, for example in selection of projects within a portfolio of a firm (Oral et al., 1991; Eilat et al., 2006).
7 duration of the multiphase R&D pipeline and which exceeds the duration of any single phase of the multiphase pipeline.
1.5.3 Variety of Outputs
Earlier literature on measurement of R&D efficiency has used a wide variety of examples of potential outputs from the R&D process, including revenue, patents and New Drug Applications (NDAs). This variety of outputs has contributed to the diversity of findings.
The advantage of DEA is that multiple outputs can be considered, so that an arbitrary choice between potentially valid output parameters does not have to be made. The selection of the multiple output parameters is undertaken using the 12 Design Principles (the model design is discussed further in
Section 4.2).
1.5.4 Types of Contributions to Knowledge
The research has produced the following contributions:
Two theoretical contributions by offering:
o additional insight into the Merger Paradox, based on the
divergence of outcomes when PAP is measured in different
ways;
o a theoretically based approach to the selection of multiple
performance measures.
A methodological contribution by introducing a novel means of
assessment of PAP, combining a longitudinal view of M&A history and
a cross-sectional view of comparative efficiency (that itself accounts for
8 the longitudinal nature of the R&D pipeline and the multiple outputs of
the R&D process).
Three empirical contributions regarding:
o scale factors for the R&D pharmaceutical pipeline;
o the statistical distribution of M&A in the pharmaceutical sector;
o differences in measurement of PAP exhibited when ROA and
ROS are used as measures of financial efficiency.
To these can be added a confirmation of a further aspect of the Merger
Paradox, namely that although M&A does not seem to be associated with higher performance to the acquiring firm, M&A value in total is associated with more efficient firms (possible reasons for this are elaborated upon later in the thesis).
9 2 Overview of Pharmaceutical Sector
2.1 Introduction
This chapter provides an overview of the research-based pharmaceutical sector, highlighting the unique characteristics of the R&D process and relating them to the research methodology.
The R&D process, which is common to all research-based firms in the sector due to the regulatory environment, is defined. The market structure of the sector that existed at the time of the analysis (2006 and the preceding decade) is then summarised. Finally, a review of the merger activity follows, including expert industry opinion on the motivations of the mergers and their consequences to the sector and the firms involved, as recorded in published reports accessed from the University databases.
2.2 R&D Processes
The pharmaceutical R&D process is unusually well-defined and recorded. This is because of the need to be confident of the safety of future compounds by undertaking tests on the human population. This has a dual advantage for this research in that well-defined R&D processes allow measurement of comparative efficiency and the metrics used in the measurement are publicly available.
Sweeny (2002: 4) provides a full summary of the pipeline:
– Discovery/Basic Research: Synthesis and Extraction – the process of
identifying new molecules with the potential to produce a desired change
in a biological system; Biological Screening and Pharmacological Testing
10 – studies to explore the pharmacological activity and therapeutic potential
of compounds.
– Preclinical Testing: Toxicology and Safety Testing – tests to determine
the potential risk a compound poses to humans and the environment
involve use of animals, tissue cultures or other test systems;
Pharmaceutical Dosage Formulation and Stability – the process of turning
an active compound into a form and strength suitable for human use.
– Regulatory Review: Application to regulatory authority to use compound
in human testing. In the USA the compound is then called an
Investigational New Drug (IND).
– Phase I Clinical Trials. Testing of a new compound in 20–80 healthy
human volunteers to determine tolerance, pharmacological effects, and
absorption, distribution, metabolism and excretion (ADME) patterns.
– Phase II Clinical Trials. Trials in 100–300 patients with the targeted
condition to determine effectiveness in treating disease or medical
condition and short-term risks.
– Phase III Clinical Trials. Trials on 1000–5000 patients to determine
clinical benefit and incidence of adverse reactions.
– Process Development for Manufacturing and Quality Control. Engineering
and manufacturing design activities to establish capacity to produce in
large volumes and to ensure stability, uniformity and overall quality.
– Bioavailability Studies. Use of healthy volunteers to show that formulation
used in trials is equivalent to product to be marketed.
11 – Regulatory Review: NDA. Application for approval to market a new drug.
In the USA this is called a NDA.
– Phase IV. Post-marketing trials to identify undetected adverse effects and
long-term morbidity and mortality profile.
This process is universally called the ‘pipeline’ and compounds move through the pipeline in stages. Measurement of a drug in the pipeline can occur at the following stages: Preclinical, Phase 1, Phase II, Phase III and Awaiting
Approval (i.e. the Phase III trial has been successively completed but
Marketing Authorisation for the NDA has yet to be given).
In practice the pipeline resembles a funnel, with many compounds entering the start and fewer emerging because the remainder fail to clear the hurdles of clinical trials. The management of the pipeline is a ‘race against time’. The patents on which the compound are originally based generally have a 30 year life, after which any company can produce the drugs on which the patent is based, in other words it becomes ‘generic’ in the lexicon of the industry. The longer a compound stays in the pipeline the shorter the exclusive manufacturing and marketing period; this leads to a considerable loss of income.
These time factors can lead to variations in approaches to management of clinical trials. A trial is focused on the use of a compound for a particular
‘indication’: treatment of a condition. Some companies choose to proceed with trials for as many indications as possible in the hope of gaining multiple marketing approvals early in the patent lifetime. However, this is also an expensive strategy because clinical trials are expensive; an alternative approach is to proceed with trials for major indications only.
12 The timing and technology of the clinical trial part of the pipeline has tended to remain relatively static, with the increase in terms of the reporting requirements being offset by advances in ICT. However the preclinical stages of the pipeline have benefited from major technological changes on two fronts:
– product technology, moving from traditional ‘small molecule’ chemical
compounds to biotechnology, where the compounds are large molecules,
derived from biological processes;
– process technology, which has allowed increased productivity in the
screening of potential drug candidates prior to clinical trials.
The latter change has implications, discussed later, for the relevance of examining R&D inputs to the process in the low productivity era.
2.3 Market Structure and Acquisition Activity
Two industry surveys, Sykes (1999) published near the start of the period of examination of M&A activity within this thesis and Hamilton (2005) published near the end, summarise the main issues facing the industry in this period.
Sykes (1999) specifically considers merger waves in the industry, correctly identifying the start of the third wave which is the focus of this study. M&A activity frequently follows waves as noted by Schoenberg & Reeves (1999) who proposed five factors that may affect acquisition activity: industry profitability, industry growth, industry concentration, capital intensity and industry deregulation; such factors have been observed in the pharmaceutical sector, as discussed below. The first M&A wave occurred in 1988–89 and led to the consolidation of a number of middling companies into top-tier firms. The second M&A wave focused on ‘mergers of equals’ or horizontal mergers intended to reduce fixed costs and increase funds available for R&D. Three
13 drivers of M&A activity were apparent: improved R&D, improved sales and marketing cost reduction, and the desire to preserve independence.
Companies were also identified as having different views on M&A, classified as merger-bent, merger-averse and merger-resistant, the last group preferring co-licensing deals to full-blown acquisition. As Sykes (1999) was going to press, the third wave commenced, with mergers involving Astra and Zeneca,
Sanofi, and Aventis. Hamilton’s (2005) study was written at the end of this merger wave that left the industry in a challenged state: “the pharmaceutical industry continues to experience problems in all aspects of its business”. R&D productivity had declined and some major drugs had been withdrawn from the market following safety concerns. At that time, the major opportunities were seen to be the emerging markets of China and India, and growing ageing and obese populations across the globe. The importance of linking the R&D strategy to commercial priorities was also emphasised, rather than focusing on exploiting new development technologies as had occurred previously.
2.4 Key Metrics
The performance of the pharmaceutical industry is measured by financial metrics similar to those used in other sectors; however, there is one particular metric that is given universal prominence in the sector, namely R&D expenditure as a proportion of revenue. For example Pharmaceutical
Research and Manufacturers of America (2010) cites the statistic on its opening ‘Key Facts’ page, and the synopsis of the sector provided by the
Association of the British Pharmaceutical Industry (2010) remarks: “Research and development lies at the heart of the pharmaceutical industry. It invests 30 per cent of its sales in research…” and then goes on to tabulate R&D as a proportion of sales over time for the sector and to compare the statistic between sectors.
14 At the firm level, the metric R&D as a percentage of sales is frequently used to rank companies by their long-term potential, on the presumption that higher
R&D expenditure leads to greater prospects of future success at the preclinical stage.
15 3 Literature Review
3.1 Introduction
The literature review for this research encompasses:
– the PAP literature from 1968 onwards, including DEA-related literature;
– the RBV, which is the theoretical basis for the selection of performance
measures;
– multidimensional measurement, especially as regards the measurement
of intangibles;
– the application of the RBV in the pharmaceutical sector;
– the small subset of the large DEA literature that considers M&A, R&D or
the pharmaceutical sector.
On analysing the literature it becomes apparent that many topics themselves are multidisciplinary. M&A in general and PAP in particular have been considered differently by different academic disciplines; performance management itself is multidisciplinary, as made clear by an extensive literature review in Neely et al. (1995). Given this, the literature review concludes with a synthesis of the various strands of literature as they relate to this thesis.
3.2 PAP
3.2.1 History
The academic literature on PAP has been accumulating for the past four decades. Weston & Mansinghka (1971) were one of the first to publish on the performance of conglomerates and were able to cite only three prior papers.
16 This paper and the references set the tone for the subsequent decades. The paper found that in a sample of 63 firms, active acquirers had lower profitability than a random sample. Of the prior papers, Reid (1968) found active acquirers scored higher on criteria related to managers’ interests than owners’ interests. Smith & Schreiner (1969) found that investment companies were better at portfolio management than corporate acquirers. Lorie & Halpern
(1970) examined if ‘deception of investors’ in the acquired firm took place but found the concerns to be unfounded with above-index returns to shareholders of the acquired firm. Therefore from very early on in the M&A literature the issues of multiple stakeholders and dubious PAP, at least for the acquiring firm as distinct from the acquired, were fully recognised.
3.2.2 Meta-Analyses
It is now recognised that PAP is an intellectual construct subject to a variety of interpretations, and for the past decade there has been an emerging sense of the need for integration of the literature seeking to unite at least some of the several theoretical perspectives. Larsson & Finkelstein (1999) seek to do this by using a structural equation model to assess how synergy realisation is affected by combination potential, organisation integration and employee resistance. Nonetheless they recognise that the synergy realisation measure is less objective than financial or accounting measures. This search for an integrative approach has also encompassed performance measures specifically: Zollo & Meier (2008) examined some but not all aspects of this construct (the ‘Performs for whom?’ question was not posed) when they undertook a meta-analysis of 87 academic articles on M&A. These papers have been subject to further analysis as discussed later. This meta-analysis revealed three broad academic disciplines: strategic management, corporate finance and organizational behaviour. The 87 studies in the meta-analysis
17 used 12 different types of performance measures. The largest group (41%) of the total used a short-term window financial event-study approach, a method that typically relies on stock market measures, as do the long-term window studies (18%) that are finding increasing application in finance journals. The next most frequent type of measure is the accounting measure (29%), which is found in the strategic management and organizational behaviour journals, whose analysis term is a matter of choice but comprises one or more years.
Other approaches attempt a more general assessment of acquisition performance, including subjective surveys and panels (14%); none of the remaining approaches total more than 7% of the total. Three broad categories of measures are therefore observed: finance (short- or long-term window,
59%), accounting (variable term, 29%) and subjective surveys (14%).
Zollo & Meier (2008) add further dimensions to their meta-analysis; firstly they consider the time dimension by using a two-way taxonomy of short and long term, acknowledging that acquisitions may be a response to immediate incentives but whose long-term effect is uncertain. In a second dimension,
Zollo & Meier (2008) also propose a three-level taxonomy: firstly, tasks involved in the acquisition, secondly the acquisition itself and thirdly the longer-term performance of the acquiring firm. Considering this three-by-two classification of measures, Zollo & Meier (2008) then provide plausible scenarios where the measures of performance may diverge: they establish that different measures may measure different aspects of the PAP construct and can be expected to diverge under certain circumstances.
Many of the 87 papers considered multiple measures and 13 examined accounting performance and one other parameter, as this thesis does; however, in no case were both accounting and operational efficiency measures for intangibles considered, which is the subject of this thesis. This
18 preference for multiple measures in the literature is an indirect endorsement of the benefits of multidimensional performance measurement and later in this review specific literature that confirms the benefit is identified.
Table 3.1 provides a chronological analysis of the Zollo & Meier (2008) papers
(this analysis was not presented in the original paper) and displays the types of measures used, with some papers considering up to three measures.
Table 3.1 Choice of Measures in Acquisition Meta-Analysis 1983–2006 Author Year First Second Third
Measure Measure Measure
Eckbo 1983 S
Jensen and Ruback 1983 S
Wansley et al. 1983 S
Buono et al. 1985 I O
Kusewitt 1985 A L
Chatterjee 1986 A S
Montgomery and Wilson 1986 V
Lubatkin 1987 L S
Ravenscraft and Scherer 1987 A
Singh and Montgomery 1987 L
Travlos 1987 S
Amit and Livnat 1988 A
Capon et al. 1988 A
Morck et al. 1988 A
Shelton 1988 S
Walsh 1988 E
Fowler and Schmidt 1989 A L
19 Walsh 1989 E
Datta and Grant 1990 O
Hunt 1990 I O
Lahey and Conn 1990 L
Seth 1990b S
Chatterjee 1991 S
Datta 1991 I O
Franks et al. 1991 S
Harris and Ravenscraft 1991 S
Harrison et al. 1991 A
Hitt et al. 1991 V
Schweiger and Denisi 1991 E
Slusky and Caves 1991 S
Chatterjee 1992 L
Chatterjee et al. 1992 S
Shanley and Correa 1992 I O
Travlos and Waegelein 1992 S
Agrawal et al. 1992 L
Cannella and Hambrick 1993 O A
Hambrick and Cannella 1993 E
Hoskisson et al. 1993 A L
Bruton et al. 1994 O
Clark and Ofek 1994 A L
Markides and Ittner 1994 S
Pennings et al. 1994 V
Berger and Ofek 1995 S
Brush 1996 A M
20 Chang 1996 A
Hitt et al. 1996 A V
Vermeulen and Barkema 1996 V
Weber 1996 I A
Anand and Singh 1997 A
Barber and Lyon 1997 L S
Covin et al. 1997 E
Hayward and Hambrick 1997 S
Holl and Kyriazis 1997 S
Krishnan et al. 1997 A
Kroll et al. 1997 S
Loughran and Vijh 1997 L
Lubatkin et al. 1997 L S
Ramaswamy 1997 A
Hitt et al. 1998 A V
Morosini et al. 1998 A
Bresman et al. 1999 I K
Capron 1999 I O
Haleblian and Finkelstein 1999 S
Larsson and Finkelstein 1999 I O
Thakor 1999 Y
Palich et al. 2000 A L S
Walker 2000 S
Ahuja and Katila 2001 N
Bergh 2001 V
Krug and Hegarty 2001 E
Beckman and Haunschild 2002 S
21 Capron and Pistre 2002 S
Hayward 2002 O S
Heron and Lie 2002 A
Seth et al. 2002 S
Carow et al. 2004 L S
DeLong and DeYoung 2004 A S
Feea and Thomas 2004 A S
Moeller et al. 2004 S
Pangarkar 2004 S
Zollo and Singh 2004 A
Harrison et al. 2005 L S
Shahrur 2005 S
Zollo and Reuer 2005 A L
Homburg and Bucerius 2006 O
Puranam et al. 2006 O
Kapoor and Lim 2007 N
Key to columns 3, 4 and 5:
I = Integration process performance; O = Overall acquisition performance; E =
Employee retention; A = Accounting performance; L = Long-term financial performance; S = Short-term financial performance; V = Acquisition survival; N
= Innovation performance; K = Knowledge transfer; Y = Systems conversion;
M = Variation in market share.
Table 3.1 shows some trends in scholarship in the examination of PAP. For the first five years, there are 2.2 papers per year and an average of 1.36 measures per paper. In the next five years output increased to 4.8 papers per
22 year but with an average of 1.2 measures per paper (i.e. adopting a one- dimensional view of merger performance). In the past ten years we see a steady 1.4 measures per paper and an average output of 3.2 papers per year.
Over time therefore the intensity of research has slightly declined but there has been a greater effort to obtain a multiparameter view.
Another recent meta-analysis of performance measures in PAP is Papadakis
& Thanos (2010), which extended work by Schoenberg (2006) and generally confirmed its results, showing merger success rates below 50%. Schoenberg
(2006) found no correlation between accounting measures, financial returns and managers’ subjective assessments, whereas Papadakis & Thanos (2010) found a correlation between accounting-based measures and managers’ subjective assessments. However, the possibility that the latter (received in a single semi-structured interview) may have been influenced by the former was not discussed in the paper. That paper considers case studies explicitly, although these can be considered a variation on a survey of subjective assessments, with a sample size of one, with the justification that each merger is so unique that any attempt at categorisation of findings into measures would risk distortion.
Another recent meta-analysis by King et al. (2004) considered whether the acquisition was by a conglomerate, whether it was related by sector, method of payment and prior experience; it also established the relative popularity of accounting measures: 29 studies using ROA, 14 using Return on Equity
(ROE) and 9 using ROS. This confirms the preference of ROA to ROE as a measure of capital efficiency because it does not depend upon the capital structure of the firm. The relative merits of ROA and ROS as a measure of financial efficiency are considered later.
23 Having established that there are four approaches to the measurement of PAP
(or three, if a case study is regarded as a subjective survey with a sample size of one) and there is little or no correlation between them, it only becomes possible to choose between them by considering the purpose for which the measures are being applied. For this thesis, one objective is to examine the
Merger Paradox, namely why M&A continue to be transacted when historically their results seem to be disappointing overall. The major references and the strengths and weakness of each method are summarised below so that judgement can be subsequently made on the most appropriate method for examination of the Merger Paradox.
3.2.3 Summary of Main Approaches
The theoretical foundation for financial performance is provided by Fama et al.’s (1969) definition of the event study and Fama’s (1970) definition of the efficient capital market hypothesis. Forty years later the validity of the hypothesis is still much discussed, however it has since become the cornerstone of modern corporate finance theory. The ‘strong’ version of the hypothesis states that prices reflect all information on a company, whether the information is public or not. If the hypothesis is true, then the ‘abnormal gains’ of share prices following a merger announcement can be considered the best possible judgment on its future performance, as expressed as the best estimation of the value created by that merger. The advantage of the method is that data are publicly available and the sample sizes are large. Several studies have suggested that mergers ‘create value’, for example
Jensen & Ruback (1983), Seth (1990b) and Singh & Montgomery (1987).
However, other studies indicate that it is the shareholders of the acquired companies who have the most consistent gains, for example
Chatterjee (1986), Datta (1991), Datta et al. (1992), Seth (1990a), Singh &
24 Montgomery (1987) and Sirower (1997). There are however two difficulties with the approach. The first is that the efficient capital markets hypothesis is still a hypothesis, especially in its strong form (the weak form states only that prices reflect public information). The second is more fundamental, namely whether a gain in wealth by the shareholders by the acquired firm (the only consistent observation) represents a genuine creation of economic value or is simply a case of overpayment, which will subsequently burden the merged firm. This raises the multistakeholder question of ‘performance for whom?’
Regarding accounting measures, these also use publicly available data and large sample sizes are available, and it is possible to monitor performance over an extended period of time. The use of accounting measures does, however, have its critics, for example it ignores risks, it treats the cost of equity and debt finance differently and the measures are historical but not forward looking, as noted by Montgomery & Wilson (1986). Notwithstanding these shortcomings, accounting measures are used by managers for decision making on the future of the firm, including decisions on acquisitions, and by financial analysts to inform forecasts that affect share prices.
The use of surveys, whether of expert panels or managers, faces the generic strengths and weaknesses of this approach. Perhaps the greatest strength is that it is possible to account in the survey for the original motives of the merger against which to assess success or failure, and Angwin (2007) stresses the importance of motive in explanation of merger decision making. Set against this is the potential for subjectivity and selectivity in survey design and tactical responses to survey questions. The case study reflects an extreme example of a survey, able to take account of the unique nuances of each acquisition and its motives, however, it is very susceptible to subjectivity and difficulties in
25 generalisation. Examples of this research approach include Haspeslagh &
Jemison (1991), Marks & Mirvis (1998) and Shanley & Correa (1992).
From the preceding discussion on the strengths and weaknesses of measurement methods and the earlier discussion on meta-analyses of studies in assessing PAP and the Merger Paradox, two important themes emerge.
The first is to understand what those who initiated the merger expected from it and the second is the need to understand what happened over a significant period of time.
3.2.4 Motive and Synergy
Brouthers et al. (1998) established that the top three motives for M&A were to
‘pursue market power’, ‘increase profitability’ and ‘marketing economies of scale’ in that order. These three motives have guided the design of this research. Firstly, the reference to ‘profitability’ suggests that accounting measures are paramount in managers’ minds, and analysis of accounting performance has been used to illuminate further the Merger Paradox
(significantly ‘profitability’ rather than ‘shareholder value’ was mentioned in the top three motives, possibly because the latter is seen as being influenced by exogenous factors); in this thesis, profitability has been measured by both
ROA and ROS. Secondly, regarding ‘market power’ and ‘marketing’, in the pharmaceutical sector this is tightly coupled with the R&D process because authorisation for particular markets or applications of compounds can only be obtained through successful completion of the clinical trial process. Therefore in this thesis, efficiency of the R&D process has been selected for examination of the PAP.
Furthermore ‘market power’ is synonymous with ‘collusive synergy’, one of three types of synergy (the other two being operational synergy and financial
26 synergy) and the RBV provides a theoretical base for the examination of synergies. Overall, synergies should be positive for an acquisition to proceed and Penrose (1959) (the earliest RBV-related paper) noted the initial presumption should be that synergies are negative unless there is a special reason otherwise. However, Rumelt (1984) notes the presence of synergies where companies diversify into areas where there are common factors.
However, the potential for synergy may not always be realised, and Angwin &
Vaara (2005) suggest there is an appreciation of the need to examine the degree of integration or connectivity with the firm.
Notwithstanding the multiple motives that are possible for a deal, Ambrosini et al. (2010) have found that acquirers that opted for a single value creation strategy, for example consolidation of costs or leverage of resources across a larger firm, experience higher PAP than those which pursue multiple strategies. In the pharmaceutical sector this has been confirmed by Higgins &
Rodriguez (2006) who noted positive financial returns to companies that sought to outsource R&D through the use of M&A to acquire technological resources.
3.2.5 Diversification Literature
The diversification literature considered synergies in more detail. Chatterjee
(1986) concluded in the Abstract: “collusive synergy is, on average, associated with the highest value. Further, the resources behind financial synergy tend to create more value than the resources behind operational synergy”.
This observation is highly pertinent to the comparison between the financial efficiency (ROA and ROS) scores, which include all three synergies, and the technical efficiency (DEA) scores, which consider operational synergies alone.
27 Examining the diversification literature more generally, there is a strong similarity with the acquisition literature. A lengthy period of research, mostly based on cross-sectional studies, has given rise to conflicting results that are now the subject of meta-analyses noting the evolution of the research. For example Martin & Sayrak (2003) note there was initially a view that there was a discount associated with diversification, there then followed a phase where it was accepted that a discount existed but that it could be accounted for by other factors, with the final conclusion that there may actually be a premium associated with diversification but there is a problem with ‘noisy proxies’ used to measure diversification, that is the principles for the measurement of diversification are being queried.
Some authors suggest that relatedness improves performance: Kitching
(1967), Elgers & Clark (1980), Kusewitt (1985), Singh & Montgomery (1987),
Shelton (1988) and Healy et al. (1997). However, as remarked previously, in some cases the ‘gains’ have included gains to target shareholders and this may simply reflect overpayment. Therefore there seems to be a consensus that some relatedness may be beneficial to the extraction of synergy, even though the earlier view that diversification lowered value is now being questioned.
Regarding cross-border diversification specifically, Seth et al. (2000) estimated total gains to be 7.6% of pre-acquisition value (i.e. including gains to target shareholders), which is comparable with the Bradley et al. (1988) figure for domestic acquisitions (i.e. there is no special advantage for cross-border acquisition) and indeed less than that observed in Eun et al. (1996), although this research did find positive total gains for cross-border deals.
28 3.3 RBV
3.3.1 Early Definition of the RBV
Although Penrose (1959) and Rumelt (1984) are now considered to be part of the RBV literature, the modern variant of the RBV was launched by Wernerfelt
(1984) who defined resources as any factor that was a strength or weakness of a firm. Some examples are given of attractive resources: Machine Capacity,
Customer Loyalty, Production Experience and Technological Leads. These particular examples have the characteristics of assets and refer to both tangible and intangible assets; these parameters are potentially measurable.
Rumelt (1984) highlighted the need to consider ‘isolating mechanisms’ that hinder the imitation of resources and cites ten factors: Causal ambiguity,
Specialised assets, Switching and search costs, Consumer and producer learning, Team embodied skills, Unique resources, Special information,
Patents and trademarks, Reputation and image, and Legal restrictions on entry.
Isolating mechanisms complicate the task of the external evaluator: it is not sufficient to identify and measure a resource, or even to compare this measurement with that of another organisation (e.g. as occurs in competitor benchmarking), but one has to anticipate the potential for imitation.
The RBV was interpreted for practitioners by Prahalad and Hamel (1990) who proposed the concept of a ‘core competency’: defined as an entity that provides access to a wide variety of markets, and `makes a significant contribution to perceived customer benefits and is difficult for a competitor to imitate.
29 3.3.2 Qualification of Resources
In the mature phase of the RBV, the perspective moved beyond proposing candidates for resources to establishing that resources had to have particular qualities if they were deliver competitive advantage. Barney (1991) proposed four essential characteristics of resources: Valuable to the company, Rare,
Imperfectly imitable and Non-substitutable (VRIN). These qualities can be used to screen potential candidates for their relevance to performance measurement.
Peteraf (1993) provided an alternative set of qualifying factors for resources when she cited the ‘four cornerstones’ to the RBV:
the heterogeneity of firms, noting that unique resources allow firms to
earn economic rent as opposed to break even;
ex-post limits to competition that limit competition for rents once
resources have been acquired;
imperfect mobility of resources, in terms of their trade;
ex-ante limits to competition, namely that there is limited competition
for resources prior to their acquisition, so as to avoid the potential
profits from being competed away by bidding for the resource.
These economically orientated factors are especially relevant to the selection of measures because they translate the qualitative concept of competitive advantage into a quantitative concept of economic rents. This is also highly relevant to the pharmaceutical industry that can be viewed as earning an economic rent on intellectual property, namely patented and approved compounds.
30 3.3.3 Dynamic RBV
In the evolution of the RBV it was becoming recognised that having a stock of resources may be necessary for competitive advantage but it was not sufficient because resources needed to be deployed: there must be a corresponding flow, or use, of the resources for some purpose, as Dierickx and Cool (1989) noted. Amit and Schoemaker (1993) provided a linkage between the emerging RBV and the earlier Industrial Organisation perspective framework, and introduced the concepts of ‘capabilities’ that were defined as the capacity to deploy resources.
The introduction of the concepts of stocks and flows into the RBV is of direct relevance to performance measurement. One can measure both a stock and a flow but care must be taken in mixing the two when building a model to evaluate efficiency.
Teece et al. (1997) highlighted the role of routines and skills in the firm in regards to the effective deployment of resources, although these factors may pose a particular challenge to measurement, especially for an external evaluator.
3.3.4 Critiques of the RBV
There have been a number of critiques of the RBV, for example Foss (1997) and Williamson (1999), and also a dialogue between Priem & Butler (2001a, b) and Barney (2001), regarding the Barney (1991) paper. The criticisms include: the RBV is tautological (instead of explaining how resources lead to competitive advantage, it assumes the point) and this makes it difficult to verify, and the RBV does not link resources to value nor does it consider the causality of how resources lead to competitive advantage.
31 This thesis addresses this weakness directly. Beginning with the observation that for a company to be in the top 50 by turnover, it must de facto be competitive, it then derives the resources that contribute to this success and uses this as a basis for a performance measurement framework (PMF).
A reassessment of the RBV was also provided by Foss & Knudsen (2003) that considered the papers of Barney (1991) and Peteraf (1993) as the core foundations of the RBV, providing the strategic management and economic bases respectively. However, these two bases were not entirely consistent, furthermore there were only two necessary conditions for sustainable competitive advantage: uncertainty and immobility. Peteraf & Barney (2003) replied, stating in the abstract that: “Unless Resource Based Theory is understood as a resource-level and efficiency orientated tool its contribution cannot be understood fully” and suggest a narrower definition of competitive advantage that focused on intra-industry advantage. This reply is entirely in sympathy with the approach taken in this paper, where the focus is on resource-level measurement to assist in the quantification of performance relative to competition within a single industry.
In conclusion this research accepts the limits to RBV proposed by its founders: its focus on intra-industry efficiency analysis. In addition, this research seeks to develop a new perspective for RBV: establishing the causality of resource possession and competitive advantage. This research is also supported by the finding in Crook et al. (2008) of a positive association of measures and performance when those measures are selected by the criteria laid out in the
RBV.
More recently, Kraaijenbrink et al. 2010 identified eight criticisms of which three were considered to merit further attention; these three were two basic
32 concepts that resource and value required a more detailed definition and there was a narrow view taken of competitive advantage.
3.3.5 Recent Retrospective on the RBV
As noted previously, the Journal of Management in September 2011 dedicated a special issue to reviewing resource based theory to commemorate its previous special issue that introduced the theory 20 years earlier; in the most recent special issue, Barney et al. (2011) considered it had reached maturity and was capable of further development to satisfy its critics.
The topic of measurement was also specifically addressed by Molloy et al.
(2011) who examined empirical tests of the RBV and found a lack of theoretical justification for the selection of the measures chosen, noting in the opening paragraph:
Resource-based theory (RBT) indicates that intangible
resources, or intangibles, underlie value creation (Penrose,
1959). A paradox of RBT is that these very resources that
underlie value creation elude examination (Barney, 2001).
Indeed, since intangibles are immaterial, scholars cannot
easily isolate, observe, or measure them (Lev, 2007). How
then are scholars to advance RBT through empirical research
that examines intangibles?
Molloy et al. (2011) propose a multidisciplinary assessment process that draws on the strengths of both economics and psychology. This thesis adopts an alternative approach of identifying factors relevant to competitive advantage that are accessible to an external evaluator.
33 3.3.6 Summary of Key Issues for Performance Measurement
Later in this thesis a set of Design Principles and a Construction Process that have been derived from the RBV is described, and as a prelude the contributions made by the main authors of the RBV to the measurement of resources is summarised in Table 3.2 following their original definition.
Table 3.2 Contributions of the Major RBV Authors to Performance
Measurement
Phase Author Contribution to Measurement
Early Wernerfelt (1984) a) Resources are the
differentiating factors
Rumelt (1984) b) Isolating mechanisms with
examples
Consolidation Barney (1991) c) VRIN tests: Valuable Rare
Imperfectly imitable Non-
substitutable
Peteraf (1993) d) Link to value and rent
generation
Dynamic Dierickx and Cool (1989) e) Importance of deployment as
opposed to possession
(prelude to process)
Amit & Schoemaker (1993) f) Capabilities (recognition of
intangible aspect to
resources)
Teece et al. (1997) g) Paths, Positions and
Processes
Reassessment Peteraf & Barney (2003) h) Efficiency perspective
34 We now discuss the topic of performance measurement in more detail, beginning with consideration of the benefits of additional multiple measures.
3.4 PMFs
3.4.1 Theoretical Benefit of Additional Information
The benefits of multiple parameter measurement for business management are considered in the next section but first we summarise the theoretical evidence. Blackwell (1951) reports that multiparameter measurement could be no worse than single-parameter measurement (although this presumed additional information was costless) but there was no view on the scale of the additional benefit. Further support comes from Holmström (1979) who considered the role of asymmetric information in a principal–agent relationship, and found that any additional information, no matter how noisy, would have a positive value. In the case of a PMF, the user of the PMF could be considered an agent, and this finding suggests that any additional information could be beneficial to either an internal or external evaluator.
This establishes a mathematical basis for the assertion that additional costless information cannot be detrimental, although it must be borne in mind that there may be a cost associated with the interpretation of the additional information and in comparative efficiency modelling additional parameters in a model can be detrimental, for example by creating the need for a larger sample.
An analogue can also be drawn with the ‘mosaic theory’ defined by Pozen
(2005: 630):
… a basic precept of intelligence gathering: Disparate items of
information, though individually of limited or no utility to their
possessor, can take on added significance when combined
35 with other items of information. Combining the items
illuminates their interrelationships and breeds analytic
synergies, so that the resulting mosaic of information is worth
more than the sum of its parts.
Having established the theoretical benefit of additional information, the next issue is how to relate this to the assessment of business performance.
3.4.2 Benefits of PMFs
PMFs are intended to provide a balanced view of the performance of the firm.
In this regard there have been positive findings on the usefulness of non- financial information to supplement conventional financial measures, for example Davis & Albright (2004) in a cross-sectional study of bank branches found better financial performance for branches implementing the Balanced
Scorecard than others. Ittner & Larcker (2003) showed that a higher ROA was associated with organisations that used PMFs than was the case with those that did not.
Ittner et al. (2003) examined financial services firms and stated (in the
Abstract):
…we find consistent evidence that firms making more
extensive use of a broad set of financial and (particularly) non-
financial measures than firms with similar strategies or value
drivers have a higher measurement system satisfaction and
stock market returns.
Banker et al. (2000) showed that including customer satisfaction as part of an incentive plan increased customer satisfaction and this led to increased revenues in a hotel chain. This finding suggests that managers can use
36 measures to influence behaviour and act to improve performance, thus establishing a causal link between PMFs and improved financial performance, as well as an association.
PMFs offer the opportunity to include measures indicating likely future financial performance as well as retrospective financial performance. Ittner & Larcker
(1998) reported a statically significant positive relation between customer satisfaction measures and future accounting performance. Furthermore they found evidence that customer satisfaction is a leading measure for financial performance, even when measured from outside the firm. There is further evidence that the benefit of leading measures is not confined to customer- related metrics. Rucci et al. (1998) found that an improvement in employee attitude translated into better customer satisfaction and revenue growth in a retail company, suggesting that the casual link extended from employee to customer to a financial measure.
3.4.3 The Balanced Scorecard and its Evolution
The Balanced Scorecard is a widely recognised form of multidimensional performance measurement proposed by Kaplan and Norton (1992, 1996) who advocated its use as a strategic management system. The Balanced
Scorecard is developed by an organisation’s management to agree the organisation’s goals, to measure and communicate progress, enrich the business plan and feedback performance to adapt strategy. The key features of the Balanced Scorecard include the need for a balanced set of measures as opposed to a single measure (four perspectives are suggested to recognise the multiple stakeholders: finance, operations, customer and employee learning) and for leading measures as well as lagging measures to be included. The relationships between measures should be expressed as a
37 performance model and the Scorecard should act as a second feedback loop, typically operating at an annual or quarterly period, to supplement the weekly or monthly operational feedback loops. Notwithstanding the remarks on feedback, the Scorecard is intended as a communication tool, intended to assist strategy deployment, not a control tool. Nonetheless some companies cascade the Scorecard down the company, assigning more detailed
Scorecards to processes and even to individuals.
The Balanced Scorecard concept is not entirely original; according to Malo
(1995), French companies have been using the tableau de bord since 1932.
Bourguignon et al. (2004) suggest however that there are differences between the two that reflect cultural differences between French and American society; certainly the Balanced Scorecard is shown to have more theoretical structure, in terms of categories and causal modelling, however this might also reflect that it is the later development, rather than any of the cultural differences suggested. It is perhaps significant that two other reports of the Balanced
Scorecard implementation in northern and southern continental Europe,
Braam & Nijssen (2004) and Papalexandris et al. (2005), respectively, did not refer to specific national cultures.
The Balanced Scorecard has evolved over time with attempts to define three phases of evolution. For example Speckbacher et al. (2003) see the first phase comprising a multidimensional framework, combining financial and non- financial measures, a second describes strategy using cause and effect relationships, and a third that implements strategy by defining objectives, plans, outcomes and incentives; this conforms quite closely to the original concepts. Lawrie & Cobbold (2004) consider the first phase as comprising the original Kaplan/Norton concepts, for example ‘balance’ and use of leading measures. The second phase is the selection of measures to be applied to
38 specific strategic objectives and the use of visual documentation of major causal relationships. The final stage involves the use of a ‘destination statement’ for the company and the development of an ‘outcome’ perspective to replace ‘financial’ and ‘customer’ perspectives, and an ‘activity’ perspective to replace ‘learning’ and ‘process’ perspectives. The last evolutions are intended to make the Scorecard more relevant to the public sector. Most of these are less concerned with the selection of measures than with their presentation and their link to change management.
Finally, in the original Scorecard there is no measure of risk and this was remarked upon by Kaplan (2010) in an interview:
If I had to say there was one thing missing that has been
revealed in the last few years, it’s that there’s nothing about
risk assessment and risk management.
Table 3.3 summarises the main lessons for performance management that arise from the Balanced Scorecard.
Table 3.3 Summary of Balanced Scorecard Concepts
Kaplan & Norton (1992) i) Need for non-financial measures
Kaplan & Norton (1992) ii) Use of leading measures
Kaplan & Norton (1996) iii) Causal links between measures
Kaplan (interview 2010) iv) Need to measure risk
The previous sources considered the benefits of particular measures with an emphasis on establishing that certain non-financial measures were leading measures of performance, however, this is not the only consideration. There is
39 also the question of populating the PMFs with measures, which is considered below.
3.4.4 Choice of Measures
Malina & Selto (2004) consider the choice of measures from the stance of management control theory and identify eight attributes, the first five being
‘design’ attributes and the remainder ‘use’ attributes. Specifically, measures should be: 1) Diverse and complementary, 2) Objective and accurate, 3)
Informative, 4) More beneficial than costly, 5) Causally related, 6) Strategic
Communication devices, 7) Incentives for improvement and 8) Supportive of improved decisions.
However there still remains the question of which measures should be chosen. Abernethy et al. (2005) proposed the building of causal performance maps to identify Key Success Factors but this is clearly difficult for an external evaluator to undertake. However, the concept of Key Success Factors seems to be closely related to Critical Success Factors that were first defined by
Rockart (1979: 85) as: “the limited number of areas in which results, if they are satisfactory, will ensure successful competitive performance for the organization”, which could be assessed by an external evaluator.
Methods for the systematic design of PMFs have been developed, even to the point of the publication of a workbook for the application of a systematic process, as described by Neely (2000). Neely at al. (2002) have developed this further into the ‘Performance Prism’. However, in these cases there was a presumption that the designers of the PMF were working with the active co- operation of the firm’s management to produce a PMF for their use; we now consider the case of PMFs designed for the use of external evaluators.
40 3.4.5 External Evaluation of Intellectual Capital
PMFs are also of use to external evaluators, especially for benchmarking purposes. Lebas & Euske (2002: 73) noted that the needs of the internal and external evaluators differ:
Performance does not have the same meaning if the evaluator
is inside or outside the organisation. The operations
management remain a black box for the outsider while the
insider operationalizes performance in cooperation with other
actors.
Where the primary audience for the non-financial measures is external, then the design of a PMF is often considered through the lens of the external reporting of a firm’s Intellectual Capital (IC). Marr et al. (2004) first consider why companies should measure IC and conclude that there is a need for more testing of the benefits, especially through longitudinal studies as opposed to cross-sectional studies (echoing similar trends in M&A research where long- term performance is an issue). Marr et al. (2004) then consider how to construct a PMF for IC, noting a sequence of definition in IC over time from
Hall (1992) where IC was considered to comprise assets and skills, through
Edvinsson & Sullivan (1996), Brooking (1996), Sveiby (1997), Roos et al.
(1997), Stewart (1997), Edvinsson & Malone (1997), Bontis et al. (1999), Lev
(2001) until Marr & Schiuma (2001) arrive at a view whereby IC is seen as comprising knowledge-based assets located either in relationships or infrastructure. This cannot be considered an entirely linear sequence of thinking (e.g. Brooking and Stewart seek focus on the financial aspects of IC).
41 The later papers have also sought to produce systematic reporting frameworks, including the IC Index (Roos et al., 1997), the IC Audit Model
(Brooking, 1996) and the Intangible Asset Model (Sveiby, 1997).
However the most commercialised framework is the Skandia Model, described by Edvinsson & Malone (1997), which divides Market Value into Financial
Capital and IC. IC is divided into Human Capital and Structural Capital, which itself comprises Customer Capital and Organisational Capital, with subsequent subdivisions of the latter. These categories can be used to group resources and act as a basis for measurement, although Roos et al. (1997) propose the aggregation of measurement into a single IC Index.
The previous literature is not sector-specific, although the use of IC frameworks in research organisations was described by Leitner & Warden
(2004) and indeed Leitner et al. (2005) experimented with the use of DEA to measure the productivity of Austrian universities, concluding it was a useful consulting tool. However, although IC frameworks provide a basis for classification of measures, they do not assist in the identification of performance measures for the external evaluation of companies in a specific sector for a specific purpose, as required by this research.
There is a major practitioner initiative led by the Enhanced Business Reporting
Consortium that has links to the accounting profession. Enhanced Business
Reporting Consortium (2006) is a framework for non-financial reporting that has been influenced by the language of the RBV, for example it suggests reporting upon ‘Competencies and Resources’. The next step is to produce industry-specific Key Performance Indicators (KPIs), although this has not yet happened. This work is being undertaken in conjunction with the World
Intellectual Capital Initiative.
42 Extended business reporting is already common in regulated industries where it is used to support comparative efficiency assessments; in the UK, the Water
Services Regulation Authority collects extensive non-financial information annually, in the form of the June returns, whose purpose is summarised in
OFWAT (2005); some of this information includes intangible parameters, such as quality of service. An example of quarterly reporting of a PMF in the
Balanced Scorecard format is the National Rail Monitor, published by the
Office of Rail Regulation; the design of the PMF is described in ORR (2004).
Also in the UK, the application of extended business reporting principles to the public sector has been supported by the National Audit Office by the issuing of the ‘FABRIC’ guidelines (Focused, Appropriate, Balanced, Robust, Integrative,
Cost Effective), as summarised in H.M. Treasury et al. (2001), and commissioning independent research (Accenture, 2009) on the design of
PMFs that comprise both financial and non-financial measures.
Separately efforts have been made to link the non-financial measurement of
IC and financial measurement. Financially, IC can be considered the Market
Value Added of the company: the difference between the Market Value and net book value of the tangible assets. Stewart (1997) has suggested a parameter termed Economic Value Added as a proxy for this, to be used as a managerial incentive, although Kramer & Pushner (1997) have questioned the evidence supporting this. Economic Value Added was intended to make adjustments to conventional accounting data to rectify some of its limitations of use as a measure, and there is no doubt that these are especially significant as regards accounting for intangibles in the context of mergers, and Canibano et al. (2000) provide a literature review on accounting for intangibles. The main issue is that internally generated intangible assets are not recognised as such, although externally acquired intangibles can be recognised as assets.
43 Thus an acquisition can lead to the recognition of assets in parameters such as ROA and result in this being a biased indicator for acquisition performance.
Boekestein (2009) has already noted the impact of M&A on the valuation of accounting value of assets in the pharmaceutical industry and this thesis explores it further.
To summarise, although various IC initiatives, whether academic or practitioner, may use the language and the concepts of the RBV to assist in their goal for standardised external reporting, they still leave open how the non-financial measures would be selected for a particular sector, for example the World Class Competitive Intelligence forum has yet to produce a draft set of KPIs for the pharmaceutical sector. The contribution of this thesis is to go beyond the use of RBV terminology and to propose a systematic approach to the design of a PMF that can be applied to any sector and then to apply it to the pharmaceutical sector specifically.
First however, we consider how the RBV has been used in the pharmaceutical sector specifically and the lesson this provides for the design of a PMF that is suitable for a comparative efficiency assessment.
3.5 Use of the RBV to Measure Performance in Pharmaceuticals
Yeoh and Roth (1999) defined the relevant resources and capabilities for a pharmaceutical firm: 1) R&D expenditures, 2) Sales force expenditure, 3)
Internal R&D efforts, 4) Therapeutic market focus, 5) Approval success and 6)
Radical innovations. This confirms the criticality of the R&D process because all factors except sales force expenditure are contributory factors to a single type of output, namely approved compounds at progressive stages in the pipeline, and sales force expenditure itself is only useful when a compound has finally emerged from the pipeline with marketing approval.
44 DeCarolis & Deeds (1999) examined the effect of stocks and flows of organizational knowledge on firm performance. Table 3.4 shows the various measures that were identified as being relevant to examining the paper’s hypotheses.
Table 3.4 Parameters and Metrics Cited in DeCarolis & Deeds (1999)
Parameter Metric
Firm performance Market value
Location Munificence of location of corporate HQ, based on factor
analysis
Alliances Number of active alliances
R&D intensity R&D expenditure as % total expenditure.
R&D pipeline Number of products at each stage contents
Patents Number of patents
Citation data Number of citations by senior personnel
Using these metrics, data from 98 firms were collected and regression models were used to test six hypotheses. The outcome is shown in Table 3.5.
Regarding the ‘supported or ‘mixed’ factors, ‘location’ refers to being based in a ‘geographic cluster’ of high performing pharmaceutical companies, as opposed to a national location, and is not relevant to this research. The remaining parameters that are found to be significant are patent citations (but not patents), the number of drugs in the pipeline and R&D intensity (i.e. expenditure). Patents themselves were not found to be linked to performance, possibly reflecting the variety of reasons for taking out a patent (including defensive reasons) and for not taking out a patent (e.g. confidentiality or
45 expense). However, citation analysis is not without its drawbacks. For example
Meyer (2001: 166) notes:
First of all, one should be aware of the general limitations of
patent citation data…citations establish only a mediated
linkage…it is not possible to derive any insight as to the
direction of potential knowledge flows.
Table 3.5 Outcomes of Hypotheses Testing in DeCarolis & Deeds (1999)
Hypothesis Outcome
Location affects performance Supported
No. alliances affects performance Not supported
R&D intensity affects performance Mixed results
No. new drugs in pipeline affects performance Supported
No. patents controlled affects performance Not supported
No. citations affects performance Supported
Furthermore, and most significantly, DeCarolis & Deeds (1999) simply summed the contents of the pipeline from Preclinical to Phase 3, which represents a significant distortion given the attrition rates between the successive stages of the pipeline, which means that compounds in the later stages have a higher value than compounds at an early stage in the pipeline.
DeCarolis (2003) also examined firm performance. The metrics used are shown in Table 3.6.
46 Table 3.6 Parameters and Metrics in DeCarolis (2003)
Parameter Metric
Firm Return on Assets performance Market-to-Book Value
Technical A measure calculated as follows: the firm issues N patents in competence a given year and within two years M patents had cited these
N patents and of these M citations, X were self-citations. The
ratio of X/N is the measure of competence.
Imitability A measure calculated as follows: the firm issues N patents in
a given year and within two years M patents had cited these
N patents and of these M citations Y were by other
companies. The ratio of Y/N is the measure imitability.
Marketing Advertising/Sales competency
Regulatory Number of new drugs per year per firm competency
Regression models were used to ascertain if there was any link between the dependent variable, firm performance, and the four other dependent variables.
The models were built with ROA and Market-to-Book Value. The results of the hypotheses testing are summarised in Table 3.7.
47 Table 3.7 Outcomes of Hypotheses from DeCarolis (2003)
Firm Performance Return on Assets Market-to-Book Value
Technical competency Positive Negative
Imitability Negative Negative
Marketing competency Not supported Positive
Regulatory competency Positive Positive
The surprising feature of the results is the difference between the results for technical competence, depending on the choice of dependent variables, with the negative correlation for Market-to-Book Value being counterintuitive.
DeCarolis (2003) provided an explanation for this by suggesting that building on the firm’s existing knowledge may be seen as developing future rigidities.
However, an alternative explanation lies in the limitations of the self-citation analysis that is the basis for measurement of technical competence; it is also noteworthy that the strength of the pipeline or its efficiency was not used to assess technical competence.
In the following chapter the lessons learnt from these pharmaceutical-related papers are used to design a PMF to support the analysis of R&D efficiency.
Prior to this however, we review the DEA literature with particular reference to acquisitions and R&D.
3.6 Relevant DEA Literature
Taveres (2002) identified 3,203 DEA publications in the period 1978 to 2002.
Only five were concerned with M&A and none were concerned with R&D or the Balanced Scorecard. In the past therefore, it seems that neither topic was of major interest, however, more recently there have been additions to the literature as discussed below.
48 We consider the DEA merger literature first and start with financial institutions.
Avrikan (1999) found acquiring banks are more efficient than target banks but efficiency was not always maintained. Sufian (2004) examined scale efficiency in the Malaysian bank industry and found both positive and negative scale effects, with smaller banks benefiting from mergers. Worthington (2004) uses
DEA to identify the determinants of merger activity in Australian cooperative deposit-taking institutions.
Another active area of DEA merger research has been hospitals. Ozgen and
Ozcan (2000) used DEA to examine scale effects in hospitals and found it to be the dominant source of efficiency improvements following mergers but technical efficiency was not affected. Ferrier & Valmanis (2004) used DEA to compare merged and non-merged hospitals and found that mergers did not have a sustained advantage in productive performance. The focus of DEA
M&A research on banks and hospitals reflects the availability of data in these institutions to undertake comparative efficiency analysis, especially as regards a large number of Decision Making Units (DMUs) with readily identifiable inputs and outputs.
R&D is less susceptible to this form of analysis, however, there has been some recent research. Linton et al. (2002) applied DEA to the portfolio optimisation of projects at Bell Laboratories. SubbaNarasimha et al. (2003) used DEA to examine the efficiency of deployment of technology knowledge; they used a composite income measure and patent-related output measure as the variables in the study.
Chen et al. (2004) in their DEA study found that R&D productivity in
Taiwanese semiconductor firms could be improved by an increase in scale.
Eilat et al. (2008) and Eilat et al. (2006) provided a multicriteria approach for
49 evaluating R&D projects within a portfolio at different stages in their lifecycle with some relevance to the computer technology industry (the papers were not explicit about the sector). Hashimoto & Haneda (2008) used DEA to measure changes in the productivity of the Japanese pharmaceutical industry over time and they found that it was monotonically decreasing. A single input was used, namely R&D expenditure, and the three outputs were patents, sales and operating profit, and they explained their choice as the best available.
DEA was also used to examine R&D productivity at a national level.
Sharma & Thomas (2008) use DEA to examine R&D between nations. Lee and Park (2005) undertook a similar international comparison for Asian economies.
DEA has also been used in conjunction with the Balanced Scorecard. Banker et al. (2004) found some trade-off in a Balanced Scorecard between ROA and non-financial measures linked to future development in the US telecommunications industry. There is a further recent Balanced Scorecard application in the field of R&D: Garcia-Valderrama et al. (2009) used a
Balanced Scorecard and DEA approach to analyse comparative performance of 90 companies in R&D in a variety of sectors in Spain.
Finally, the only truly comparable DEA study to this research is
Hashimoto & Haneda (2008) in which the DEA outputs are actually financial parameters only loosely associated with one DEA input: R&D expenditure.
3.7 Synthesis
The focus of this thesis on the measurement principles and parameters that determine PAP is fully contemporary for the M&A field. The potential for diversity includes:
50 – Differing motives for the M&A deal that may include market power,
financial performance or efficiency.
– A choice of different parameters for measuring progress towards the
same objective, for example ROA or ROS, often made without a clear
explanation.
– Potential bias of some metrics for measuring crucial aspects of
performance, for example accounting conventions not to record internally
generated intangible assets.
– Different approaches for the measurement of efficiency, including the
processes included and the inputs and outputs considered, which can be
especially challenging when their outputs are intangible and their
production has an element of uncertainty, as is the case with R&D.
In the face of this diversity, the approach of this thesis has been to adopt a rigorous approach to the selection of measures used in the analysis. The rigour began with first establishing that multiple measures are beneficial and having done so establishing a theoretical basis for the selection of measures for the purposes of external evaluation. The verification by Crook et al. (2008) of a link between a firm’s financial performance and its resources as identified by the RBV is confirmation of the suitability in principle of the RBV as a means for the identification of non-financial measures to populate a PMF.
Furthermore, because the RBV has accumulated literature over two decades it also provides a source of practical guidance for the design of a set of principles for the selection of measures, as well as examples of measures that have been used in the pharmaceutical sector in the past. We now describe a practical set of RBV-based principles for the design of a PMF.
51 4 Design of PMF
4.1 Introduction
This chapter of the thesis develops a systematic approach to the design of a
PMF:
– The first step is to develop a general set of Design Principles, derived
from the RBV and the Balanced Scorecard literature reviewed in Chapter
3, which can be used to develop a PMF in any sector in a structured way.
– The second step is to apply the Design Principles to the pharmaceutical
sector in order to produce a PMF; this is compared with prior literature to
demonstrate that the Design Principles represent an advance on previous
thinking.
– Finally, a reduced set of measures is selected that is suited to the
application of DEA to the pharmaceutical R&D process, along with an
explanation of why that process was selected.
We begin with describing the creation of the Design Principles from a theoretical basis.
4.2 PMF Design Principles
The lessons for the design of a PMF have been annotated (a) to (h) in Table
3.2 and (i) to (iv) in Table 3.3 with reference to academic authors. The sources in these 2 tables have provided 12 lessons and they are arranged in Table 4.1 by topic and they retain the original table references.
52 Table 4.1 Design Principles
Category No. Principle Ref.
Scope and Structure
Stakeholders 1 Recognise requirements of major stakeholders (i)
of the firm (e.g. the firm’s owner and customers)
with non-financial measures.
Leading 2 Include at least one leading measure (typically (ii) measures non-financial).
Risk 3 Risk exposure should be measured. (iv)
Resources and Barriers
Differences 4 The PMF should measure the critical few (a)
differences between firms.
VRIN 5 Measures should be on how resources are (c)
VRIN.
Link to value 6 Measures should consider how resources are (d)
linked to value creation or economic rent.
Intangibles 7 Capabilities, or intangible resources, should be (f)
measured where possible.
Barriers 8 Barriers to imitation that affect the value of a (b)
resource should be measured.
Processes and Positions
Deployment 9 Deployment of resources, through processes, (e)
should be measured.
Dynamic links 10 Causal links between measures should be (iii)
identified to provide a time dimension.
Static positions 11 Static positions are candidates for (g)
measurement.
Efficiency and Benchmarking
Benchmarking 12 Efficiency should be measured, recognising (h)
prior paths of development, for benchmarking.
53 The relatively few references to measurement within the core RBV texts allowed these references to be included without selectivity; the scarcity of references to measurement is perhaps unsurprising given the comments of
Kraaijenbrink et al. 2010 that resources require closer definition. Given this ambiguity, the issue of measurement has been to the forefront.
In the case of the Balanced Scorecard, the issue was the opposite, namely there exists a superfluity of commentary on measure selection, albeit for an evaluator with access to the internal working of the firm. The approach here was therefore to identify the main principles of the Scorecard (the balanced provided by multiple perspectives and leading and lagging indicators) and the need to understand the inter-relationships between the measures; to this was added Kaplan’s view on the main omission, namely risk.
The grouping of the twelve principles into four headings was not driven by the literature itself but by identifying major association. However there are also secondary associations and these are shown graphically in Figure 4.1. This figure summarises the relationship between the principles and the major references, while also showing the primary and secondary linkages (the secondary linkages are shown with a dotted line) and the key references from the literature review (whose attachment is marked by a small circular symbol).
54 Figure 4.1 Summary of Design Principles
55 Having derived the 12 Design Principles from the academic literature, without regard to the availability of information, it is now necessary to confirm this availability, if the principles are to be of practical use to an external evaluator.
4.3 Availability of Information
4.3.1 Practicality
The availability of information is discussed below, by considering each heading in turn.
4.3.2 Scope and Structure
Stakeholders
The identification of stakeholders should be straightforward but there may be complications, for example different classes of shareholders may have different interests (especially as regards exposure to risk, for example ordinary versus preference shareholders) and these might need to be recognised.
The customers of a company are usually evident to an external evaluator, although complications can arise; this is especially so in the pharmaceutical industry where the eventual consumer of a product, the patient, may well differ from the purchaser. However, the actual identities of the consumer and purchaser are typically known. Other common stakeholders include employees and society as a whole, whose interests may be represented by regulators.
A good understanding of stakeholders is essential because they are the medium through which competition is felt, for example competition for
56 customers, employees or scarce resources. Ultimately ‘value’ derives from satisfying stakeholder requirements (primarily those of the paying customer) and later principles require the link to value to be established.
Leading Measures
Leading measures of performance often include measures of stakeholder satisfaction, both customer and employee; the challenge is the collection of information, although polling is one option that has been used by investors where there is a matter for concern.
Other leading measures might include the results of a product development programme or the emergence of substitutes.
Risk
Assessment of risk exposure for an external evaluator is more straightforward for publicly quoted companies because there are disclosure requirements, for example in the USA, the Annual 10K form required by the Securities and
Exchange Commission gives disclosure of major risks in ‘Item 1A – Risk
Factors’. Otherwise some form of due diligence is required, for example to detect overdependence on certain products or customers.
4.3.3 Resources and Barriers
Critical Differences
The next step is fundamental to the entire PMF design, and it requires the identification of Resources and this requires some sector knowledge even if specific company knowledge is only incompletely known to the external evaluator. Nonetheless, some company-specific knowledge is also essential because the RBV focuses on the differences between firms competing in the
57 same market. For public companies there are minimum disclosure requirements. For companies quoted in the USA, Form 10-K, discloses tangible assets in ‘Item 2 – Properties’ and ‘Item 1 - Business’ provides a comprehensive list of products, markets and customers. Similar disclosures are made in other jurisdictions.
Fortunately, in publicly quoted firms the management often see that it is in their interest to publicly disclose some of the unique strengths of their companies beyond the minimum required in annual reports thus enabling establishment of a fuller picture of resources.
VRIN (Valuable, Rare, Imperfectly imitable and Non-substitutable)
The VRIN tests are used to screen potential candidates for measurement.
Although a potential parameter may seem valuable it might not be a useful addition to PMF, for example if it is easily imitated.
Link to Value
Even if a resource is VRIN, it may not merit measurement if its link to value creation is obscure, for example the possession of prestige premises. The most direct link to economic value derives from the satisfaction of customers’ requirements and the fulfilment of a product or service.
Intangibles
Identification of intangible resources may be less straightforward, although details of patents and trademarks are in the public domain. The most elusive intangible is knowledge, especially tacit knowledge.
58 Barriers
In some cases, barriers to imitation of resources will be sufficiently important to be an object of measurement in their own right. For example, for a technology company the remaining life on a patent is a crucial measure.
Similarly, where brand strength is crucial, then monitoring of major brands by independent valuation may be merited.
4.3.4 Processes and Positions
Deployment
Resources are deployed through processes and these offer an opportunity for measurement; they need to be identified and placed in the context of the wider supply chain outside the firm. Order fulfilment processes are externally visible through the delivery of products and services, although product and market development processes may be less visible (fortunately this is not the case in the pharmaceutical sector). In selecting process measures, the external evaluator, as Lebas & Euske (2002) noted, is less interested in ‘action variables’ used for process management but in the ‘critical few’ variables, to use the phrase of Murray & Richardson (2000): parameters of process performance that are crucial to the success of the organisation.
Dynamic Links
The presence of multiple measures in a PMF can be a source of confusion if their dependencies are not understood. Identifying associations between measures has an additional potential advantage of understanding positive or negative interdependencies that arise from underlying business processes.
59 Static Positions
In many cases performance will not depend on accumulation of past performance into the present position. Generally, the interest of an external evaluator will be in the firm’s position relative to competitors and there is a rich source of academic and practitioner literature on competitor intelligence, for example Lackman et al. (2000) provide a comparison of 16 competitor intelligence functions. The widespread presence of such functions is evidence not only of the feasibility of collecting information on a company from an external perspective but also demonstrates the usefulness of the information obtained.
4.3.5 Efficiency
Benchmarking
Efficiency measurement and benchmarking requires information on both inputs and outputs of a process that are often expressed as a ratio. Financial ratios are a type of efficiency analysis where information is readily available, although ratios formed with at least one non-financial parameter can also be illuminating (e.g. sales per employee). Trends in efficiency ratios are often of equal interest.
Another approach to efficiency analysis, adopted in this thesis, is comparative efficiency in which the efficiency of a company relative to its peers is measured.
60 4.4 Application of Design Principles to the Pharmaceutical Sector
4.4.1 Demonstration of Use
Having established the practicality of using the Design Principles to populate a
PMF in general, there remains the task to apply it to the pharmaceutical sector and demonstrate that the results are superior to less systematic methods.
The application of the Design Principles to the pharmaceutical sector is considered below and the resulting PMF presented in the following section.
4.4.2 Scope and Structure
Stakeholders
The following stakeholders were identified:
– The shareholders or owner of the company, who are concerned with
financial performance. Earnings per Share is of prime interest and is
linked to share price through the Price Earnings ratio for the sector.
– The customer stakeholders are unusually complex with different
customers having different priorities, for example national health
authorities and hospitals will be relatively more cost-conscious, whereas
the patient and physicians will be concerned about the efficacy of the
formulation. Notwithstanding the structural complexity of the stakeholders,
satisfaction however can be measured by market share in particular
therapeutic categories, reflecting the efficacy of the treatment and its
affordability, although loss of share can occur through lapse of patents,
allowing low-cost competition to erode market share, or the existence of
product substitutes in therapeutic categories that themselves may not be
patent-protected.
61 – The employees are natural stakeholders but are a diverse group with
some groups such as research workers being crucial. The turnover of this
group is of special interest, as is the value-added per employee in total
(the difference in revenue and costs excluding payroll), to understand the
average financial contribution of each employee.
– Society is a key stakeholder for the pharmaceutical industry with a need
for a supply of new and better formulations and also expectations that
dangerous formulations will not be released. This stakeholder is
represented by the regulatory authorities (e.g. the Food and Drug
Administration in the USA) who serve adverse reports and notices when
society is judged to be at risk.
Four internal and external stakeholders have therefore been identified, to which is added Process. Although not a stakeholder in its own right, process efficiency is a necessary condition for satisfaction of the other stakeholders and all stakeholders will have an interest in it.
Leading Measures
The R&D process is an obvious source of leading measures and it might be thought that a patent with a 30-year life is also a potential candidate for a measure, given the legal protection against imitation. However, the number of patents was not used because of criticisms in the academic literature, including the pharmaceutical R&D literature, of the usefulness of this measure.
To amplify, DeCarolis & Deeds (1999) established the lack of correlation between patents and a firm’s performance and identified patent citations as an alternative. However, the counterintuitive results in DeCarolis (2003) regarding the negative correlation between technical competence and Market-to-Book
Value raises questions over the use of citation analysis, so this variant of
62 patent analysis was not used, especially given the criticism of citation analysis by Meyer (2001).
Given the rejection of patents as a measure, the measurement of outputs therefore focuses on the number of compounds that have passed the hurdle of approval by the regulatory authorities.
Risk
The decisions of the regulatory authorities also represent major risks to the company and merit measurement; fortunately adverse reports on products and facilities are publicised.
There is also a further negative factor that requires recognition, namely litigation. Major pharmaceutical companies are usually engaged in large litigation suits and of these some are opportunistic.
4.4.3 Resources and Barriers
Differences
The primary differentiation for a research-based pharmaceutical company is current and future product portfolio, the latter being represented by its R&D pipeline.
The benefit of the current portfolio is visible through the ROS and market share and the latter can be seen as a measure of competence in marketing.
Regarding production facilities, the role of this tangible resource is to ensure continuity of supply.
63 VRIN, Link to Value & Intangibles
The previous discussion on focusing on differences between the firms highlights the importance of seeking measures that focus on the current and future product portfolio, as evidenced by ROS and the R&D pipeline respectively. Measures in these areas already pass the tests on VRIN, establishing a link to value and give full recognition to intangibles, as do any measures related to research employees, for example retention measures.
However production facilities are usually not rare and production can be outsourced; manufacturing competence is expressed as the avoidance of regulatory censure while ensuring continuity of supply.
Barriers
The prime isolating mechanism for a research-based pharmaceutical company is the patent. However these have a fixed life (30 years) so the barrier to imitation erodes with time. Products can also be grouped into therapeutic categories that address particular medical conditions; the specificity of the action of a drug is an isolating mechanism because a drug does not compete against drugs as a whole, but only against those that treat the same condition.
4.4.4 Processes and Positions
Deployment and Dynamic Links
For the pharmaceutical sector, key processes include:
Discovering new drugs (i.e. New Chemical Entities) that successfully
pass through preclinical development.
64 Rapidly and successfully progressing them through the clinical trials
process to maximise the useful patent life. This depends on the
avoidance of unnecessary delay in communication with regulatory
authorities.
Communicating with and satisfying regulatory agencies.
Marketing the drugs in as many markets as possible and identifying as
many indications of a drug for different medical conditions as possible.
These topics have largely been addressed in the preceding discussion except for the dynamic aspect: the avoidance of regulatory delay while a compound is in the R&D pipeline and the need to address not only the number of drugs produced but also the number of clinical trials for multiple indications. Both these additional factors need to be addressed by the PMF.
Static Positions
The static position of a company in a market is best observed by its revenue in general and by its market share in therapeutic categories specifically.
However, the prime concern of the industry is the erosion of a static position by patent expiry.
4.4.5 Efficiency
Benchmarking
Regarding efficiency ratios, the prime concern in the industry is availability of funds to invest in R&D with R&D as a proportion of sales being seen as a prime metric of a firm’s commitment to the future. Consequently ROS to fund the expenditure on R&D is seen as the key financial metric.
65 4.5 Proposed Pharmaceutical PMF
4.5.1 Outcome of Design Principles
The outcome of the application of the Design Principles is shown in Table 4.2.
The measures have been grouped by stakeholder, rather than linked back to the category of the Design Principles by which they were created, so as to present a view on the coverage and balance of measures by the stakeholders identified in Subsection 4.4.2.
Table 4.2 Proposed Pharmaceutical PMF
Owners Customers Employees Processes Regulatory
►Return ►Market ►Research ►# compounds in ► Patent life on Sales share by employees as pipeline, by stage lost to
►Earning therapeutic % total ►# clinical trials submission per Share category ►Value added in pipeline, by process (-ve)
►Litigation ►Future loss per employee stage ► # adverse liability (- of sales to ►Employee ►R&D/Sales drug reports ve) therapeutic turnover (-ve) ►Marketing/Sales (-ve)
substitutes ►Attrition rates in ► # adverse
(-ve) pipeline (-ve) manufacturing
►Future loss reports
of sales to (-ve)
expiring
patents (-ve)
66 Although Table 4.2 represents an illustration of the concept of the design principles, the question arises whether the scorecard is a useful set of measures for measurement of pharmaceutical performance or an advance on contemporary practice.
An opportunity for comparison was provided by Stankevicien & Svidersk
(2010). The scorecard was designed by a German subsidiary of an Italian company described thus (Stankevicien & Svidersk, 2010: 242): “The Italian
Group enjoys an outstanding reputation worldwide as an efficient and reliable partner. This applies both to the development of new drugs and to the communication of scientific insights”. The company designed the scorecard shown in Table 4.3.
Table 4.3 Scorecard for Pharmaceutical Company
High Performance Systematic Stakeholder Excellence in
Organization Execution & Service Financial
Implementation of Excellence Performance
System
Requirements
Employee Audit Business Risk
satisfaction recommendation improvement management
rate Implementation rate score
Training score Management Expense
compliance Information satisfaction spending
rate sharing score rate control
Successful job External Performance
rotation rating score score
Target
achievement
67 rate
Comparison of the two tables suggest that the Design Principles have led to a scorecard whose measures focus on the competitive position of the company rather than process compliance and improvement, which are the natural concerns of the management of a subsidiary.
4.5.2 Use of Negative Indicators
A feature of the PMF produced by the Design Principles is its inclusion of large numbers of negative indicators; this is uncommon in the Scorecard literature in which emphasis is placed upon a monitoring rate of strategic deployment.
However, the RBV stresses the importance of barriers to competition and the lessening of barriers has an important influence on the value of resources and these represent a negative indicator for future performance. In recognition of this, one of the measures listed in Table 4.2 relates to future sales lost to patent expiry.
Litigation risks over consumer harm and patent infringements are also major topics of interest and represent a leading measure of performance.
The inclusion of negative indicators can be seen as fully contemporary in its recognition of risk to the business, as noted in an interview with Kaplan, held in 2010, as discussed previously.
4.5.3 Choice of Financial Measures
ROS has been selected rather than ROA. The first reason concerns the limitation of ROA in this sector: the denominator in ROA without an acquisition history will include only tangible assets, as internally generated intangible
68 assets are not shown according to accounting conventions. The measure is therefore of limited relevance in a sector where profitability depends on earning a return on intangible assets.
ROS by contrast allows the profitability of the company to be expressed as a percentage of revenue, allowing for inter-firm comparison, and shows the funds available for investment in R&D, with R&D as a percentage of revenue being an accepted approach for comparisons of research-based pharmaceutical companies, for example Grabowski et al. (2002).
Despite the advantages in ROS, this research does consider ROA in parallel to the ROS to further examine the relative merit of the two measures.
4.6 Focus on R&D and Adaptation to DEA
Comparative efficiency is measured at the level of the DMU however this still leaves some discretion whether to measure the comparative efficiency of an entire firm or a particular process within the firm. If one opts for the entire firm, then there is the possibility of blurring the analysis by aggregation of many factors, whereas the application of the analysis to one element of a firm, such as an individual process, risks overlooking important interdependencies between processes that can affect performance. In this thesis the decision has been to apply the analysis of comparative efficiency at the level of a major business process within the firm, at a high-enough level so that the impact on the firm’s competitiveness can be understood, while isolating the intangible production processes from the other parts of the firm.
The selection of the R&D process for examination was made because a necessary but not sufficient condition for the success of an ethical pharmaceutical firm is its R&D core competence, as evidenced by an efficient
69 R&D pipeline. This pipeline is the process that provides deployment of capabilities that generate resources that can be most easily converted to income. These resources are the multiple outputs of the pipeline: compounds at progressive stages of the pipeline that meet the VRIN tests required by the
RBV by virtue of patent protection. This protection ensures an economic rent is earned once a compound has progressed to the end of the pipeline and is granted marketing approval. Once marketing approval has been granted, additional competences become important, especially marketing to maximise revenue during the period of patent protection and manufacturing, to ensure continuity of supply. However before marketing-related resources can become valuable there must be compounds to produce and hence the primary focus for an evaluator is the efficiency of the R&D pipeline.
Once the R&D process has been selected for examination of comparative efficiency, there remains the task of choosing from Table 4.2 a mix of input and output parameters relevant to the R&D process. The candidates are:
Inputs: Research employees, % R&D expenditure.
Outputs: Compounds, Trials, Attrition rates (negative).
These are the measures used in the analysis, except the number of research employees is not available so the total number of employees has been used instead. This does have the advantage of recognising the indirect input of non- research employees to the operation of the R&D pipeline; the total number of staff may also be relevant in examining the link to acquisition history because acquisitions are generally followed by staff reductions.
The attrition rates were not included in the DEA model for two reasons, namely the difficulty in collecting information on failed trials over the historical
70 period and the difficulty of including outputs with a negative influence on productivity into a DEA model (one option is to include a reciprocal of the output but this itself poses a problem where there is zero attrition).
4.7 Conclusion
This chapter has outlined a set of measurement principles for an external evaluator to measure the performance of a pharmaceutical company with a comparative efficiency technique. The focus is on the R&D process and inputs and outputs have been selected on the basis of a RBV-derived process and comparison with prior literature.
Future chapters will develop these principles into a DEA model that will enable the modelling of comparative efficiency from an external perspective.
71 5 Research Methodology
5.1 Introduction
This chapter describes the research methodology that has been used to examine the association between a pharmaceutical firm’s history of acquisitions and the efficiency of its R&D process.
The first step was to consider the range of possible efficiency and productivity analysis approaches, and the decision to opt for a DEA approach is explained.
Several DEA models have been developed within the literature and the selected model types are explained, along with their mathematical formulation.
Weight restrictions have been applied to the DEA model and their implications are considered.
Returns to scale for the pharmaceutical R&D process has been investigated and tested as a preliminary step. The primary focus of the research however is the association between efficiency and acquisition history. Statistics on acquisition history were collected and their distributions analysed prior to the development of a typology to support the subsequent analysis.
Various hypotheses have been proposed and their testing requires an approach that avoids the difficulty that arises from statistical testing of DEA scores directly; this is then defined, followed by the application to the testing of the scale hypothesis and the acquisition-related hypotheses. Finally, the treatment of outliers is also discussed.
5.2 Efficiency and Productivity Analysis Approaches
Efficiency and productivity measurement (the terms are not synonymous but have been used interchangeably in the literature) has a long history and one
72 of the earliest measures is unit cost. The seminal author in the field of measurement of productive efficiency is Farrell (1957) who introduced the distinction between technical and allocative efficiencies. Despite this early work, the analysis of R&D efficiency has been largely confined to examining the technical aspect through the consideration of unit costs: dividing an output measure (typically number of approved compounds) by an input measure
(typically expenditure). However, unit costs cannot measure the efficiency of a productive unit with multiple inputs or outputs. Unit costs were not employed in this research because the R&D process has multiple outputs.
An approach which is able to consider multiple inputs and outputs is to use
‘parametric’ methods, whose characteristic is an assumption of the form of the production function relating inputs to outputs. The simplest parametric method is Ordinary Least Squares regression and more sophisticated approaches include Stochastic Frontier Analysis (Kumbhakar, 1988) that separates the effect of noise (e.g. from measurement error) from variation in efficiency. If the form of the production function is known, then parametric analysis confers several advantages for econometric analysis (as noted by Cubbin &
Tzanidakis, 1998). However in this case, the form of the production function linking R&D expenditures (inputs) to the outputs of the process is not known.
Therefore a parametric approach was not adopted for the efficiency analysis in this research.
Given the presence of multiple inputs and outputs and the lack of knowledge of the production function, the natural choice for the efficiency measurement is
DEA, which is not a statistical or econometric technique but has its origins in
Operations Research.
73 The formulation of the DEA models is described below, once the mathematical notation used in the thesis has been defined. DEA has been selected for the evaluation of R&D productivity because it is able to analyse multiple inputs and outputs and does not require the production function to be specified. In this section we specify several DEA models that:
– Recognise the longitudinal nature of pharmaceutical R&D, namely that it
may take many years of R&D to produce a measurable output. Therefore
R&D over several years has been considered.
– Have the potential to consider both financial and non-financial outputs,
while investigating the differing restrictions on the weights given to these
two different types of outputs.
We now define the structure of the model in more detail and in particular how the longitudinal aspect of the pipeline is addressed.
Following this, the design of the acquisition typology, used to analyse acquisition history, is summarised.
5.3 Summary of DEA Model
5.3.1 Orientation of Model
There is a distinction between an input-orientated model (that examines efficiency from the viewpoint of minimising resources used for a given level of output) and an output-orientated model (that examines efficiency from the viewpoint of maximising output for a given level of resources consumed). The output-orientated model is more relevant to pharmaceutical R&D because the generic strategy within the industry is to ‘speculate to accumulate’: to spend
74 available surplus on R&D in the hope of discovering new compounds to secure the future of the company.
5.3.2 Inputs and Outputs
For all the models used, the inputs included were R&D (current), that is expenditure in 2006, and R&D (historic), that is expenditure in the previous five years (2001–2005), expressed in US dollars; the definition of this is likely to be relatively standard between firms and include all the operating costs of a firm’s R&D facilities, including staff. Although R&D expenditure is the prime monetary input to the creation of preclinical compounds and the progression of compounds through the R&D pipeline, it is necessary to consider a sufficient number of years of expenditure to relate the input to outputs in all stages of the pipeline.
However, other factors are also required for pharmaceutical R&D, for example activity on dealing with regulatory agencies, staff recruitment and retention, and raising finance and so on. These may need to be reflected in the inputs in some way, therefore number of staff was also considered an additional input in one of the DEA models.
In the DEA model used for association with acquisitions, the outputs are:
– the number of compounds in Preclinical phase (i.e. yet to gain approval
for clinical trials to commence);
– the number of compounds in Phase 1 clinical trials;
– the number of compounds in Phase 2 clinical trials, having passed
Phase 1;
75 – the number of compounds in Phase 3 clinical trials, having passed
Phase 2;
– the number of compounds awaiting approval for marketing, having
passed Phase 3.
An alternative model considering the number of clinical trials was also considered for comparison, although after examination of the trends in the ratio of trials to compounds it was considered less representative of R&D process efficiency and more reflective of marketing strategy.
5.4 Longitudinal Dimension
Pharmaceutical compounds take time to develop, so R&D expenditure is unlikely to produce a preclinical compound in the same year because it takes many years for a drug to move from discovery to the market place.
DiMasi & Grabowski (2007) provide a contemporary summary of the times and costs involved (see table below).
Table 5.1 Time and Costs of R&D Development
Testing Phase Duration (months) Monthly cost ($m, 2005)
Preclinical 52.0 1.15
Phase I 19.5 1.66
Phase II 29.3 1.08
Phase III 32.9 1.38
Six years, or 72 months, of R&D expenditure has been collected for this thesis. This period not only covers the majority of the duration of the pipeline
(which is 133.7 months) but also is greater than any phase of the pipeline (a maximum of 52 months), so therefore the inputs can be related to each output.
76 Table 5.1 also shows that monthly costs per phase are comparable between phases and therefore the measured input (R&D expenditure) for 72 months should be representative of the cost required to support the pipeline production process as a whole. Further evidence to support this is given by the analysis of the descriptive statistic for R&D expenditure which shows that
R&D expenditure as a proportion of sales varies little year-on-year.
Including R&D expenditure relating to years prior to 2001 could be misleading because this expenditure would relate to the discovery and preclinical stages of compounds now in the later stages of the pipeline. The advent of combinatorial chemistry for screening in the decade prior to the year 2000 had a major impact on the productivity of this stage in the pipeline as noted by
Sweeny (2002: 10)
This was a major rate-limiting step in developing new drugs
and has seen remarkable increases in productivity over the
past ten years or so through the use of combinatorial chemistry
linked to high throughput screening.
5.5 Definition of Notation
5.5.1 MCT
There are two MCT that are relevant to this research: the arithmetic mean and the median; the first is pertinent to the parametric test for the difference between two means (used in the statistical testing) and the second is relevant to the non-parametric test. Both measures have been used for statistical testing of the mean for the parametric tests and the median for the non- parametric test. Where appropriate, the acronym ‘MCT’ (i.e. Measure of
77 Central Tendency) has been used in the terminology below; the choice of
MCT was determined by the test used: parametric or non-parametric.
5.5.2 Normalisation Factor
In quantifying merger history for a particular firm, it is necessary to establish a normalisation factor for the firm because firms vary considerably in size. There are various options for the selection of a normalisation factor. A natural choice of the normalisation factor would be the assets of the company because an acquisition involves an expansion of the asset base; however this presents some technical difficulties because many assets were acquired in times of different asset prices and have been subject to varying depreciation policies.
Furthermore the figure does not include many of the self-generated intangible assets on which pharmaceutical companies earn a return (these criticisms affect the usefulness of the ROA as a performance measure despite its popularity). Because pharmaceutical companies earn a return on intangible assets through the sale of medicines whose price reflects the value of those assets, the normalisation factor could be based on the profitability of the firm; however, profitability, being the difference between cost of sales and revenue, can vary considerably between years. Revenue itself is also a commonly used scaling factor in industry analysis and is often used to compare R&D intensity between firms, however it can be affected by competitive conditions not related to the underlying scale of assets or processes of the firm. Therefore cost of sales, which is usually of comparable magnitude to revenue and has an underlying proportionality to the maintenance of the tangible and intangible assets, has been selected. The sum of the deal value over the period in question has been divided by the annual cost of sales of the surviving company at the end of the period. This allows the significance of the merger history to be expressed independently of the size of the resulting company.
78 In the notation below the acronym ‘NDV’ (i.e. Normalised Deal Value) is used to denote the cumulative values of deals of a firm over the analysis period
(chosen to include an entire merger wave and economic cycle) divided by the revenue scaling factor.
5.5.3 Algebraic Notation for Efficiency Analysis
Tables 5.2 sets out the algebraic notation used in the DEA modelling.
(Subentry Tables 5.3 to 5.5 define further algebraic notation; the notation is first defined completely so that the equations can then be considered without interruption of further definition of terms).
79 Table 5.2 Algebraic Notation for DEA Models
Symbol General Meaning Specific Meaning # Elements
in Variable
S # output measures Number of compounds at 1
different stages of pipeline
(value = 5)
M # input measures Number of input measures; 1
(these are R&D current, R&D
historic, and staff numbers) i.e.
value = 3)
N # DMUs Number of firms in sample after 1
exclusion of two outliers (value
= 48) yik Value of output Compounds or clinical trials for 5 × 48
measure i (i = 1, …, each firm (i=1, Preclinical; i=2,
s) for DMU k (k = 1, Phase 1; i=3, Phase 2; i=4,
…, n) Phase 3; i=5, Awaiting
Approval) xjk Value (≥ 0) of input R&D spend for each firm (j=1, 3 × 48
measure j (j = 1, …, Current; j=2, Historical) and
m) for DMU k (k = staff numbers where used (j=3)
1, …, n) ui Weight (> 0) of Weight given to each compound 5
output measure i (i or clinical trial for each firm
= 1, …, s) vj Weight (> 0) of Weight given to R&D current, 2
80 input measure j (j = R&D historic or staff numbers
1, …, m) for each firm d'k Optimal objective Reciprocal of technical output 48
function value for efficiency for CRS model
each DMU d’’k Optimal objective Reciprocal of technical output 48
function value for efficiency for VRS model
each DMU
ηk Relative technical Technical output efficiency 48
output efficiency of score for each firm from CRS
each DMU (CRS) model
θk Relative pure Pure technical output efficiency 48
technical output score for each firm from VRS
efficiency of each model
DMU (VRS) rk Financial efficiency ROS or ROA or SOA for each 48
of each DMU firm
5.5.4 Form of DEA Equations
The equations defining DEA can be expressed in three forms. The first is the original Fractional Programming form in which the efficiency of each DMU is expressed in the form of a ratio. These equations can be restated in Linear
Programming form of which there are two variants: one is called the Multiplier form, and there is also a dual form termed the Envelopment form. In this thesis the Multiplier form is used.
81 5.5.5 Algebraic Notation for Examination of Returns to Scale
Table 5.3 sets out the algebraic notation used in the examination of returns to scale based on the use of the average of the Current R&D expenditure and the Historic R&D expenditure as an R&D-specific scale factor.
Table 5.3 Algebraic Notation for Examining Returns to Scale
Symbol Specific Meaning #
Elements
in Variable ek Scale efficiency of each DMU, Ek = ηk / θk, for k = 1 … n 48 wk Mean of Current and Historic R&D expenditure of each 48
DMU wl Mean of wk with below-median scale efficiency, ek 1 wh Mean of wk with above-median scale efficiency, ek 1
Var(w)l Variance of average of Current R&D and Historic R&D
expenditure of DMUs with below-median scale
efficiency, ek
Var(w)h Variance of average Current R&D and Historic R&D
expenditure of DMUs with above-median scale
efficiency, ek
5.5.6 Algebraic Notation for Classification of Acquisition History
Table 5.4 sets out the algebraic notation used in the examination of acquisition history.
82 Table 5.4 Algebraic Notation for Classification of Acquisition History
Symbol Specific Meaning # elements in
Variable
Ak Sum of all deal values for DMUk, for k = 1 … n 48
Bk Sum of cross-border deal values for DMUk for k = 1 48
… n
Ck Sum of cross-sector deal values for DMUk for k = 1 48
… n
Fk Annual Cost of Sales for DMUk for k = 1 … n 48
ak NDVs for DMUk for k = 1 … n 48
bk NDVs of cross-border deals for DMUk for k = 1 … n 48
ck NDVs of cross-sector deal for DMUk for k = 1 … n 48
al MCT of ak for those DMUs with below-median θk 1
bl MCT of bk for those DMUs with below-median θk 1
cl MCT of ck for those DMUs with below-median θk 1
ah MCT of ak for those DMUs with above-median θk 1
bh MCT of bk for those DMUs with above-median θk 1
ch MCT of ck for those DMUs with above-median θk 1
a'l MCT of ak for those DMUs with below-median rk 1
b’l MCT of bk for those DMUs with below-median rk 1
c’l MCT of ck for those DMUs with below-median rk 1
a’h MCT of ak for those DMUs with above-median rk 1
b’h MCT of bk for those DMUs with above-median rk 1
c’h MCT of ck for those DMUs with above-median rk 1
Var(a)l Variance of ak for those DMUs with below-median θk 1
Var(b)l Variance of bk for those DMUs with below-median θk 1
Var(c)l Variance of ck for those DMUs with below-median θk 1
Var(a)h Variance of ak for those DMUs with above-median 1
θk
83 Var(b)h Variance of bk for those DMUs with above-median 1
θk
Var(c)h Variance of ck for those DMUs with above-median θk 1
Var(a)'l Variance of ak for those DMUs with below-median rk 1
Var(b)’l Variance of bk for those DMUs with below-median rk 1
Var(c)’l Variance of ck for those DMUs with below-median rk 1
Var(a)’h Variance of ak for those DMUs with above-median rk 1
Var(b)’h Variance of bk for those DMUs with above-median rk 1
Var(c)’h Variance of ck for those DMUs with above-median rk 1
5.5.7 Algebraic Notation for Statistical Testing
It is necessary to avoid the statistical testing of DEA scores directly because they are not independent observations. To avoid the testing of scores the approach adopted has been to use the DEA parameter to divide the population into two groups on the basis of the DEA scores: one a group with a below-median DEA efficiency and the other group with an above-median DEA efficiency. Table 5.5 defines the notation used to describe the variables used in the statistical tests employed for hypothesis testing using this approach.
84 Table 5.5 Algebraic Notation for Hypothesis Testing
Symbol General Meaning
µl Mean of below-median group
µh Mean of above-median group
2 σl Variance of below-median group
2 σh Variance of above-median group
πl Number in below-median group
πh Number in above-median group
T t test statistic
Rl Sum of ranks for lower-median group
Rh Sum of ranks for upper-median group
Ul U-test statistic for lower-median group
Uh U-test statistic for upper-median group
U U = min (Ul, Uh)
Z z-test statistic
5.6 DEA
DEA was proposed by Charnes et al. (1978) in a paper that began “This paper is concerned with developing measures of ‘decision making efficiency’’” and coined the term DMU; in this research DMU refers to 1 of 48 pharmaceutical firms. The paper then defined what has since been termed the Charnes,
Cooper & Rhodes (CCR) model in the Fractional Programming form and the two Linear Programming forms.
There is a choice to be made between the input- and output-orientated form of the DEA model. The output-orientated model maximises the outputs for a
85 given level of input whereas the input-orientated model minimises the input for a given level of output. Because a pharmaceutical firm ‘speculates to accumulate’ and commits surplus resource to R&D to maximise R&D output, the output-orientated form is the more appropriate and has been selected.
An output-orientated form of the CCR model is described below. Also described is a later model that was developed to accommodate VRS.
5.7 CCR Model
The original CCR model calculates an efficiency score for each DMU, based on its ratio of multiple outputs to its multiple inputs, weighted so as to maximise its efficiency, subject to constraints that all the DMUs have an efficiency less than or equal to unity. DEA can therefore be seen as an extended formulation of unit cost analysis.
The output-oriented form of the CCR model is summarised below in the multiplier form that establishes the relative efficiency for the DMU under consideration: DMU0 (as opposed to an absolute efficiency based on technical standards).
s
Min d’0 = Σ vj xio Eq. (1) i = 1 s
Subject to: Σ uj yio = 1 i = 1 s m
Σ uj yik ≤ Σ vj xik k = 1 ... n
j = 1 j =1 ui > 0 i = 1 … s vj > 0 j = 1 … m
where the subscript ‘0’ refers to the element of any variable relating to
the DMU under analysis.
86 A Linear Programming equation has three aspects: an objective function to be optimised, a set of variables and some constraints. In the equation above, the optimisation relates to the efficiency of the DMU under analysis and the value of d0’ is the reciprocal of the technical output efficiency score, η0. The variables are the weights and the operation of the DEA optimisation algorithm assigns weights to each DMU that maximises its efficiency (although later in this section we discuss the inclusion of weight restrictions). The constraints ensure that the efficiencies of all the DMUs are less than or equal to unity with the chosen weights. The optimised choice of weights will allow the efficiency of each DMU to be the highest possible and at least one DMU will lie on the
‘efficiency frontier’ (i.e. be technically efficient whereas typically some others are relatively inefficient) and have a score of d’k = 1.
A basic property of the CCR model is that it assumes there are no economies or diseconomies of scale: it assumes CRS. Subsequently alternative DEA models have been developed, one of which allows VRS and is described below. In this research the CCR model is used only to establish economies of scale by comparing the CRS efficiency scores with the VRS efficiency scores.
5.8 BCC Model
The model used for the acquisition hypothesis testing is the Banker, Charnes and Cooper (BCC) model, defined by Banker et al. (1984). This is an extension to the CCR model (Eq. 1) that accommodates VRS through the addition of an additional free scalar variable. The output-orientated BCC model is defined below:
m
Min d’’j’ = Σ vj xio – v0 Eq. (2)
87 j = 1 s
Subject to: Σ uj yio = 1
j = 1 s m
Σ uj yik ≤ Σ vj xik + v0 k = 1… n
j = 1 j=1 ui > 0 i = 1 … s vj > 0 j = 1 … m
where the subscript ‘0’ refers to the element of any variable relating to
the DMU under analysis.
It can be seen that the BCC model has an additional term, v0. The value of d’’0 is the reciprocal of the technical output efficiency score, θk. The firms on the efficiency frontier will have a value of d’’k = 1.
5.9 Output Weight Restrictions
Weight restrictions are a means of incorporating subjective judgements into a
DEA model and may be of two types: absolute or relative. Without restrictions it is possible for DEA to generate weights that conflict with the judgements of the DMUs’ decision makers; an example from this research would be for a lower weight to be given to a compound in a late stage in the R&D pipeline, compared with a weight in an earlier stage in the pipeline, when the latter must still face cost and uncertainty to moving forward to subsequent stages (i.e. uj-1 < uj,forj=2…5).
Weight restrictions improve the credibility of the model in the eyes of decision makers but the subjective judgements involved must be defended. In this case, simple relative output weight restrictions have been applied along the lines indicated by Wong & Beasley (1990). Specifically four restrictions have been applied:
88 uj ≥ uj-1 forj=2...5, Eq.(3)
where j = 5 represents the output relating to the compound at
the final stage of development
These output weight restrictions are applied to both the VRS and CRS models.
5.10 Input Weight Restrictions
It has been found that the current CRS and VRS models tend to often assign zero weights to one or the other of the two inputs in the current models:
Historical R&D (i.e. the annual historical average) and Current R&D expenditure, whose magnitudes are similar. Input weight restrictions have been added that limit the extent to which either input can be reduced to zero to recognise that both inputs are required to produce the outputs. The restrictions are:
u1 ≥ 0.5 u 2 and u2 ≥ 0.5 u 1 Eqs (4 & 5)
These input weight restrictions are applied to both the VRS and CRS models and have the effect of ensuring both R&D inputs have a material effect on the model, while allowing each input to be up to two times as significant as the other. However, the introduction of the input weights given in Eqs (4 and 5) had only a minor observed effect on the efficiency scores and given this minor effect further variations on the arbitrary 0.5 factor in Eqs 4 & 5 were not made.
5.11 Returns to Scale
The research afforded an opportunity for a fresh examination of returns to scale in pharmaceutical R&D using DEA. Banker et al. (1984) suggested the possibility of the use of the CCR model to relate returns to scale to the size of
89 the firm and introduced the concept of scale efficiency as defined in Eq. 6, which is calculated by reference to the technical output efficiency produced by the CRS model (Eq. 1) and pure technical output efficiency produced by a
VRS model (Eq. 2), or in mathematical form:
ek = ηk ÷ θk k=1…n Eq.(6)
where the subscript k represents the DMU under analysis
The scale efficiencies do not in themselves provide a test for whether there are returns to scale. A statistical test for investigating this is described below, following a description of the analysis of acquisition history.
5.12 Design and Population of Acquisition Typology
5.12.1 Identification of Acquisitions
The Thomson One Banker database has been used to identify all acquisitions over a ten-year period with a deal value exceeding $100 million, where the acquiring company was in the North American Industry Code (NAIC) for
“Medicinal and Botanical Manufacturing, Pharmaceutical Preparation
Manufacturing, In-Vitro Diagnostic Substance Manufacturing, Biological
Product (except Diagnostic) Manufacturing”.
Corporate acquisitions are one of many means by which a firm may purchase technological or marketing resources. Licences are preferred for minor acquisitions of resources which represent a stream of activity that would be undetected by the research methodology. From 1998 to 2002 the average value of a licensing deal was $84.5m (Pharmaventures, 2003). To exclude alternative means of resource acquisition from the study, for this thesis a $100 million threshold was set on the M&A analysis to ensure that the effects of
90 alternative lower-value means of acquiring resources, such as licensing, would not affect the analysis of the scale of historical resource acquisition for each firm.
For the M&A analysis, a period from 1993 to 2005 was selected; this spans at least an entire merger wave (commencing in 1993) and also approximates to the length of a typical economic cycle. The year in which R&D outputs are measured is the following year, namely 2006, to avoid the M&A activity affecting the collection of data on R&D outputs. The selection criteria used to identify deals in the Thomson One database are given in Table 5.6.
Table 5.6 Selection Criteria
Search Term Scope
Acquirer NAIC Medicinal and Botanical Manufacturing
or Pharmaceutical Preparation Manufacturing
Acquirer Ultimate Parent In-Vitro Diagnostic Substance
Primary NAIC (Code) Manufacturing
Biological Product (except Diagnostic)
Manufacturing
Date Unconditional 01/01/1993 to 31/12/2005
Ranking Value inc. Net 100 upwards
Debt of Target ($Mil)
Per cent of Shares 51 upwards
Owned after Transaction
91 The searching process led to 591 acquisitions which met the criteria.
5.12.2 Classification of Acquisitions
The initial classification of acquisitions is as follows:
– The name of the acquirer.
– The size of the acquisition is measured by the ‘Ranking Deal Value’: a
parameter used by the Thomson database to identify the value of the
acquisition (in essence the amount paid by the acquirer after adjustment
for debt).
– A binary value indicating whether the acquirer and the target are in the
same country (i.e. a cross-border acquisition).
– A binary value indicating if the target is in the same NAIC code as the
acquirer, or not (i.e. a cross-sector acquisition).
This classification makes it possible to calculate the sum of the deal values in total for each named acquirer, and in addition the sum of the cross-border deal values and cross-sector deals for each acquirer, as required by the research methodology.
The identification of cross-border and cross-sector acquisitions is amplified further below.
5.12.3 Identification of Cross-border and Cross-sector Acquisitions
The simple binary classification of cross-border and cross-sector deals has the advantage of data being readily available and reflects the additional complexity involved in acquiring a company in a different nation, for example dealing with different jurisdictions, the additional complexity of accounting and
92 control procedures and more costly logistics. It might be argued that the classification neglects:
– Geographic distance, however in practice information flows are now
instantaneous.
– Language differences, however English is now the lingua franca for the
pharmaceutical industry.
– Cultural differences. This is a significant issue, for example a US/UK or a
Swiss/German acquisition is likely to encounter fewer cultural obstacles
than say, a US/Japanese acquisition. Although techniques exist to
measure cultural distances between nations their application is
complicated by firm-specific differences and a resolution of these
differences is impractical.
The simple binary classification is then used to select out the cross-border acquisitions from the total and these are then linked to the major pharmaceutical companies.
Regarding cross-sector acquisitions, pharmaceutical companies have a number of acquisition options available to them:
– To remain within their current field but to acquire emerging technologies,
for example a traditional chemically based company choosing to acquire
more contemporary biotechnology expertise.
– To acquire closely associated non-pharmacological technology, for
example acquiring delivery devices for the administration of medicines
being produced by the company.
93 – To diversify into related businesses within the same value-chain, for
example by acquiring the means of distribution of its products.
– To undertake unrelated diversifications.
This research has adopted a relatively simple classification of diversification: a binary classification that would classify the first of these four options as undiversified and the remaining options as a full diversification.
The primary reason for opting for a simple classification relates to sample size.
Although the various types of diversification listed above could be subjectively assessed and categorised reliably, the size of the sample in each category would be very small, especially because pharmaceutical companies have generally had an aversion to making acquisitions outside their core business.
It would have therefore been difficult to obtain statistically significant results with a more granular classification of cross-border and cross-sector deals.
5.12.4 Linkage to Major Pharmaceutical Companies
The top 50 pharmaceutical companies were identified based on their health- care revenue generated during 2006, as recorded in Pharmalive (2007). Of these, details of the R&D pipeline were available for 48 that were the focus of the research and termed the ‘major’ companies below. Access to the database on which the report was based was also purchased and specific queries were resolved with the company. It is possible to check specific items of data on the database against public records.
Not all of the major companies had undertaken acquisitions and not all acquisitions were undertaken by these companies. However, out of the 591 acquisitions identified by the Thomson database in the sector during the 10
94 years’ history, 140 related to the major pharmaceutical companies, of which 64 were cross-border and 29 were cross-sector acquisitions.
The resulting analysis measures M&A history where an acquiring company has since been acquired by a ‘surviving’ major company, (e.g. the acquisitions allocated to AstraZeneca include those for both Astra and Zeneca).
5.13 Analysis of M&A History
The collection of data of acquisitions and the boundaries set on the size of deal and the periods considered are described in Chapter 6 along with the acquisition typology. The key terms in the merger typology are defined in
Table 5.7.
Once populated, the deal values for all deals, cross-border deals and cross- sector deals, are summed to arrive at the SDV values, Ak,Bk,Ck respectively, for k = 1 … 48, for the three cases. These are then divided by the annual cost of sales for the firms, Dk, to arrive at the NDV, ak, bk, ck, for k = 1 … 48, for each of three cases. Both the SDV and the NDV that are applied in the statistical test approach are described in the following section.
95 Table 5.7 Key Terms for Acquisition Typology
Term Meaning
Cross- A deal where the acquirer and the acquired have headquarters in border different nations
Deals An acquisition which results in majority control of the acquired firm
by the acquiring firm
Deal The ‘ranking deal value’ of the deal as specified on the Thomson
Value One database. Broadly, this is the amount paid for the acquisition
used in ‘ranking’ the deal in league tables
Firm One of the Top 50 pharmaceutical companies existing in 2006 that
has not been eliminated as an outlier
SDV The sum of the deal values for an acquirer
NDV SDV divided by the annual cost of sales of that firm
Cross- A deal where the acquired company does not have a sector pharmaceutical Standard Industry Code
5.14 Statistical Test Approach
Traditional statistical testing, for example differences in mean DEA scores, (by parametric methods) is problematic because the scores are not independent.
Grosskopf (1996) discussed approaches to the resolution of this problem; however, the approach in this thesis has been to circumvent the problem entirely by applying statistical tests to independent variables that were not themselves derived from DEA.
The hypothesis testing was initially undertaken using a test of significance based on a variation of the Student t test, a test proposed by Welch (1947),
96 suitable for testing two samples with unequal variances (the variances were calculated and shown to be unequal). This test is used to establish the confidence with which the difference between the two group means could be considered significant (a one-tailed or two-tailed test was used as appropriate, reflecting the phrasing of the null and the alternative hypotheses).
The t test depends on the calculation of a test statistic, T, given by:
2 2 -0.5 To = (µh - µl) (σl /πl + σh /πh) Eq. (7)
where the terms are as defined in Table 5.5 and the subscript ‘0’ refers
to the element of any variable relating to the DMU under analysis.
In this thesis, n = 48 and is even, so πl = πh = n/2. The T statistic is used to calculate the p-value by reference to the integral of the probability density function of the Student’s t distribution. That p-value is then used in the hypothesis testing and compared with thresholds, as defined by
Bowerman et al. (2011: 360).
In order to undertake the test for returns to scale for the R&D process the total set firms were divided into two subgroups: one with below-median scale efficiency and the other with above-median scale efficiency. For each group the mean of R&D current (x1k, for k = 1 … n) for firms in the pair of groups was calculated, wl and wh, for the below-median and above-median groups respectively. In Eq. (7), µl to µh were set equal to wl and wh respectively and
2 2 the variances σl and σh were set equal to Var(w)l and Var(w)h respectively.
In order to test the association between acquisition history and pure technical output efficiency a similar process was followed, namely θk,fork=1…nwas used to divide the firms into two subgroups with lower-median and upper-
97 median values of θk. For each group, the statistics ak, bk, ck, namely the NDVs for all deals, cross-border and cross-sector respectively, were used to calculate the mean values for the lower-median and upper-median groups. In
Eq. (7):
the value of µl was set equal to al, bl, cl;
the value of µh was set equal to ah, bh, ch;
2 the variance σl and were set equal to Var(a)l, Var(b)l, and Var(c)l;
2 the variance σh was set equal to Var(a)h, Var(b)h, and Var(c)h.
The testing of the ROS-related hypotheses followed an identical form to the testing of pure technical efficiency.
There was a second stage in the significance testing. A visual inspection of the distribution of the variables xk ak, bk and ck indicated a non-normal distribution. This was confirmed by the use of an Anderson–Darling test
(1952). An additional significance test was therefore undertaken using a non- parametric test, namely the Mann–Whitney U test (Mann & Whitney, 1947), shown in Eq. (8).
Ul = πl πh + 0.5 πl (πl + 1) – Rl Eq. (8)
Uh = πl πh + 0.5 πh (πh + 1) – Rh
U = min (Ul, Uh)
To apply the test, the ranks of the parameters of the two groups were calculated and the sums of the ranks in the two subgroups (Rl and Rh in Eq. 8) were calculated for each; Ul and Uh are then calculated and the lower value is used for the statistical test (in this case πl = πh = n/2 because n is even).
98 For samples where there are over 20 items in each group (i.e. n/2 > 20), as is the case in this research, the U statistic can be considered to be normally distributed and the z statistic can be calculated as shown in Eq. (9):
-0.5 z = (U – 0.5 πl πh ) (πl πh (πl + πh + 1)/12) Eq. (9)
where the terms are as defined in Table 5.5 and the subscript ‘0’ refers
to the element of any variable relating to the DMU under analysis.
The z statistic is then used to generate a p-value by reference to the normal distribution and the p-value is used for hypothesis testing as described earlier for the parametric tests.
Both the parametric Welch test and the non-parametric Mann–Whitney test were used for testing because some assumptions were violated in both
(normality in the first and equal variances in the second); Zimmerman (1998) compares the approaches under the violation of assumptions. The paper found that non-parametric methods may not be acceptable substitutes for parametric methods when parametric assumptions are violated, and the approach in this thesis has therefore been to use both methods.
5.15 Incomplete Data
There were two cases of companies that appeared in the ‘Top 50’ list of global pharmaceutical companies, where it was not possible to establish their pharmaceutical R&D expenditure. Therefore these companies were excluded from the analysis, leaving a sample of 48 companies.
99 6 Data and Descriptive Statistics
6.1 Introduction
This chapter describes:
the data used as inputs and outputs of the DEA models;
the results from the DEA models that are used to examine the
association with acquisition history;
the financial data ROS, ROA and SOA are also described;
the calculation of the acquisition history statistics, SDV and NDV.
The DEA model data are presented in tables given in Appendix D and the full acquisition data are presented in Appendix E, where the acquisition typology is populated for deals in aggregate, cross-border deals and cross-sector deals.
Having described the data, descriptive statistics are provided for:
the R&D process, including the relation between scale efficiency and
the R&D scale variable and the distribution of the R&D scale variable;
the parameters used as DEA outputs: an assessment of the
differences between the numbers of clinical trials and numbers of
compounds;
acquisitions in the sector, to establish the relation between size and
frequency, and the NDV of major firms.
100 In some cases this analysis produces empirical findings in its own right and in the remaining cases it is preparatory work for the statistical testing of hypotheses. This chapter concludes with a summary of the main findings.
6.2 Available Input Data for the DEA Models
The available input data (not all the data are used in all the DEA models) are presented in Table D.1. The third column presents R&D expenditure for 2006 in $million. The fourth column shows a calculated figure for the historical expenditure that is expressed in 2006 currency values; the calculation principles are described in Appendix A and in summary comprise adjusting historical data for R&D expenditure back to 2001 for inflation and the ratio of
R&D expenditure to sales. The fifth column is a figure for staff levels, based on a five year average of staffing levels, and the data for the calculations are shown in Appendix B.
Table D.1 also defines an abbreviated code for each company and the codes are used in later tables.
6.3 Available Output Data for the DEA Models
There are two possible output sets for the DEA models: the number of compounds produced for approval and the number of applications of those compounds either within the clinical trial process or awaiting approval. The numbers of compounds are shown in Table D.2.
Table D.3 shows the number of clinical trials at each stage of the pipeline and the ratio of trials to compounds at each stage. The ratios of trials to compounds at different stages of the pipeline can be used to draw inferences on which parameter is the most appropriate to use as an output for a DEA
101 model to examine R&D efficiency, although both models are built for both cases and their results compared.
6.4 Comparing CRS and VRS Efficiency Scores
The input data in Table D.1 relating to R&D only and the output data in
Table D.2 were applied to both a VRS and a CRS output-orientated model.
The resultant VRS and CRS efficiency scores, by firm, are shown in the second and third columns of Table D.4. The fourth column gives the calculated scale efficiency and the fifth column the natural logarithm of that parameter. The sixth column shows the natural logarithm of the mean of the two R&D inputs.
Table D.5 shows similar information to Table D.4 except that the outputs of the
DEA model are the number of clinical trials as opposed to the number of compounds.
6.5 DEA Model Results: Comparing Input Assumptions
The previous results of the VRS DEA models did not include staff as an input.
Table D.6 shows a comparison of R&D expenditure (current and historical) as inputs and then additionally with staff numbers as an additional input. In both cases, the outputs were the number of compounds (as opposed to the number of clinical trials).
102 6.6 Financial Efficiency Data
The ROS, ROA and SOA (formed by dividing ROS by ROA) are shown in
Table D.7 for each firm.
6.7 Acquisition and NDV Data
The full acquisition data, extracted from the Thomson One Banker database, are provided in Appendix E. It comprises details on the deals and permits analysis by size, nation of acquirer and acquired, and the sector of the acquirer and acquired. Table D.8 gives the total value number of acquisitions for each company: SDV, the normalising factor (cost of sales and NDV), and the resulting NDV for each company.
Descriptive statistics on acquisitions and NDVs are provided later.
6.8 Association of M&A and Technical Efficiency
The bisection of the acquisition history statistics into two subgroups, below- median and above-median pure technical efficiency is shown in Tables D.9,
D.10 and D.11.
Tables D.9 and D.10 list NDV and they show the bisection based on two different DEA models, and Table D.11 shows the bisection of SDV using the base model. In the table headings ‘
Table D.9 shows the bisection based on the VRS technical efficiencies that are calculated using the number of compounds as an output and R&D expenditure alone as an input. This is the base model used for comparisons with alternatives.
103 Table D.10 shows the bisection based on the VRS technical efficiencies using the number of compounds as an output and both R&D expenditure as well as staff numbers as inputs.
The outcome of the hypothesis testing is when the two different DEA models are later compared.
6.9 Association of M&A and Financial Efficiency
The bisection of the acquisition history statistics into two subgroups on the basis of financial efficiency is shown in Tables D.12, D.13 and D.14.
Table D.12 shows the bisection based on ROS, Table D.13 shows the bisection for ROA and Table D.14 shows the bisection based on SOA.
6.10 Descriptive Statistics of the DEA Model Outputs
The initial choice of outputs for the measurement framework was made with reference to the RBV and the consideration of what constitutes a resource of the firm. On this basis the number of compounds and the number of clinical trials within the R&D pipeline were both identified as resources because they represented potential future revenue. However, it was not possible to use the
RBV further to make a choice between these options as to which would be the better parameter to use to measure R&D productivity. It may be the case that a higher ratio of trials to compounds indicates that a firm is producing more productive compounds, or it may simply be the case that the firm is choosing to incur higher development costs for higher eventual return from the compounds they have available.
104 It is, however, possible to examine the data of the proportions of compounds to trials and draw inferences. Table 6.1 shows the mean and standard deviation of the ratio of clinical trials to compounds through the pipeline.
Table 6.1 Descriptive Statistics of Ratio of Trials to Compounds
Phase PreClinical Phase I Phase II Phase III Awaiting Approval
Mean 1.16 1.26 1.56 1.65 1.35
Std Dev 0.25 0.36 0.54 0.60 0.58
Table 6.1 shows a small ratio in the Preclinical and Phase 1 stages, where the interaction between the compound and a human is first examined. After this, the ratio increases as does the variance, as the compound enters subsequent more expensive phases. The results for the ‘Awaiting Approval’ may indicate that where large numbers of trials have been commissioned then some of the more speculative trials did not produce the intended results.
This qualitative reasoning was supported by a series of paired two-tailed t tests undertaken between the ratios of compounds at each stage of the pipeline on whether they were taken from the same sample (see Table 6.2).
Table 6.2 t test on Ratios of Trials to Compounds in Successive Phases
Phase Preclinical/Phase 1 Phase I/II Phase II/III Phase III/AA p-value 34% 0.003% 38% 0.7%
These statistics are consistent with the hypothesis that the safety of a compound on humans is first confirmed in Preclinical and Phase 1, with no statistical significant difference between the ratios of compounds per trial. A commercial decision is then made as to whether to incur considerable
105 expense funding multiple trials in Phase II and Phase III, or to adopt a more cautious approach by restricting the more expensive clinical trials to the most promising indications. Finally there seems to be confirmation that some degree of caution in commissioning trials is justified because the ratio of trials to compounds going forward for eventual approval drops back, and there is a statically significant difference between the ratio for Phase 3 and Awaiting
Approval (by contrast to the absence of a difference between Phase II and phase III).
Given this analysis, the primary examination of R&D efficiency has been undertaken by considering the number of compounds as the output of the
R&D process, as opposed to the number of clinical trials.
6.11 Descriptive Statistics of the R&D Process
Figure 6.1 shows a plot of the logarithm of the scale efficiency and the logarithm of the average of the current and historic annual R&D expenditure for the 48 firms. The graph shows that there is an apparent linear relationship between the two variables when the output of the DEA model is considered to be the number of compounds. A regression line is also given as a visual aid, although no formal regression was undertaken.
106 Figure 6.1. Graph of Scale Efficiency Versus Mean R&D (Compounds as
Output)
0 3 4 5 6 7 8 9 -0.5 ) y c
n -1 e i c i f f E
-1.5 e l a c S ( -2 n l
-2.5
-3 ln (mean R&D)
Figure 6.2 shows an equivalent graph but with the number of trials as an output. The linear relationship is still apparent but the degree of scatter is higher. This could imply that there are fewer exogenous influences on production when the output of the R&D processes is taken to be the number of compounds as opposed to the number of trials.
Figure 6.2 Graph of Scale Efficiency Versus Mean R&D (Trials as Output)
0 3 4 5 6 7 8 9 -0.2 )
y -0.4 c n e i c i -0.6 f f E
e l
a -0.8 c S (
n l -1
-1.2
-1.4 ln (mean R&D)
107 The graphs above do not in themselves test for scale, although they are strongly suggestive that scale effects exist. To test for scale, the model with the number of compounds as an output was used. The scale efficiencies of the firms was used to bisect the sample into two subgroups of above-median and below-median efficiency. The mean of the R&D expenditure of the firms in the two subgroups was then tested for a statistically significant difference.
The distribution of R&D expenditure for the two groups is shown in Figure 6.3
Figure 6.3 Distribution of R&D Expenditure of Above- and Below-Median
Efficiency
Visually it is apparent that the subgroups have markedly different distributions however formal hypothesis testing is problematic. The subgroup with below- median scale efficiency does not have a normal distribution and therefore the assumptions of the Welch test, which requires normally distributed data, have
108 been violated. The non-parametric Mann–Whitney test can be used, however this requires equal variances in the subgroups and in the sample above the variances differ by 40%. Therefore a triangulation approach has been employed, using both a parametric approach and a non-parametric approach, with the results being compared (as explained further in Section 7.2).
6.12 Descriptive Statistics of M&A
Park et al. (2010) have asserted that the distribution of size of M&A versus frequency follows a power law: the logarithm of the size of an occurrence is proportional to the increase of the frequency. This research offered the opportunity to examine this claim in the context of the pharmaceutical sector.
The distribution of M&A deal size in the acquisitions for the 510 pharmaceutical acquisitions in the full sample was examined by plotting logarithm of value by frequency, as shown in Figure 6.4.
Figure 6.4 Frequency of M&A Deals by Value
Distribution of Deal Value
1000
100 s l
a -0.3888x
e y = 182.71e D
2
. R = 0.9042 o N 10
1 1 2 3 4 5 6 7 8 910111213 Ln (Deal Value)
109 The results generally are consistent with a power law (as indicated by the linear trendline on log/log scales), however it must be stressed the evidence is not conclusive because other distributions can show a similar relationship (see
Appendix C). It can be seen that for very large acquisitions the linear relationship breaks down, with a disproportionate number of ‘mega-mergers’
(four) in the top category in Figure 6.4 and a dearth of numbers of mergers in the categories immediately below (although this effect might depend on the threshold of adjacent categories used in that figure). Of the four mega- mergers two were undertaken by Pfizer, one by Sanofi-Aventis and one by
GlaxoSmithKline. Pfizers’ ROS was close to mean, GlaxoSmithKline’s near the top of the range and Sanofi-Aventis’s relatively poor. This is consistent with a diversity of motive for the mega-mergers, with Pfizer acting as a profitable predator and the Sanofi-Aventis deal being a defensive merger
(which was actually encouraged by the French government to preserve a
French major pharmaceutical company2).
NDV has been used as a measure of acquisition history for the individual firm and the distribution of this parameter was also examined. Park et al. (2010) note that when examining acquisition frequency a Poisson distribution is commonly assumed, and given acquisitions are discrete events this is not implausible.
A goodness-of-fit test for a Poisson distribution was performed and the expected versus observed values showed a higher than expected number of zeros (shown in column 1, Figure 6.4). If the 11 zeros are removed and the square root of the remaining NDV’s is subject to a normality test (McCullagh &
Nelder, 1989: 196, explain that a feature of the Poisson distribution is that the
2 http://news.bbc.co.uk/1/hi/business/3658639.stm
110 square root of the distribution may approximate normal), then the resulting p- value is close to 5%, as Figure 6.5 confirms.
Figure 6.5 Normality Test for Square Root of NDV less Zeros
The establishment that the square root of NDV is normally distributed once zeros are removed is an interesting finding that suggests that the use of Zero-
Inflated Poisson models may have some future application in analysing merger behaviour or undertaking statistical tests, however this was not taken further in this thesis.
111 6.13 Summary
The differences in measures of mean NDV, which will be tested in the next chapter, are summarised in Table 6.3 for technical efficiency and Table 6.4 for financial efficiency.
Table 6.3 Summary of Association of Acquisition History with Technical
Efficiency
Table ak ak bk bk ck ck
Ref. θk
NDV, DEA Base Model 6.9 0.74 0.91 0.33 0.21 0.06 0.03
NDV, DEA Staff Model 6.10 0.72 0.92 0.36 0.18 0.06 0.03
SDV, DEA Base Model 6.11 4345 19174 1056 4442 472 744
Table 6.4 Summary of Association of Acquisition History with Financial
Efficiency
Table a’k a’k b’k b’k c’k c’k
Ref. θk
NDV, ROS 6.12 0.50 1.15 0.13 0.41 0.06 0.02
ROA 6.13 0.70 0.95 0.28 0.26 0.044 0.041
SOA 6.14 1.03 0.62 0.31 0.23 0.02 0.07
The MCTs of the subgroups with above and below efficiencies are now subject to statistical testing (see Chapter 7).
112 7 Hypothesis Testing
7.1 Scope of Hypothesis Testing
7.1.1 Classical Hypothesis Testing Model
The hypotheses have been divided into five sets, as explained in Section 1.4.
The hypotheses are tested using p-values derived from parametric and non- parametric tests, using what Sheshkin (2011: 57) terms the ‘classical hypothesis testing model’, which is a fusion of the work of Fisher (1925) who proposed the concept of the null hypothesis and Neyman & Pearson (1933) who developed the concept of the alternative hypothesis. There were substantial differences between Fisher (1925) and Neyman and Pearson
(1933) and the hybrid development (that is now termed the ‘classical hypothesis testing model’) was developed in subsequent text books; Lehmann
(1993) examines the consistency of the hybrid and considers there is statistical consistency.
A comprehensive contemporary textbook, Sheskin (2011: 58) outlines the key terms of the classical model, beginning with the null hypothesis which is defined as “a statement of no effect or no difference’. In this thesis, the null hypothesis is therefore taken to correspond to the status of the literature prior to the research. Having defined the null hypothesis, Sheskin (2011: 58) then defines the second key concept: “The alternative hypothesis, on the other hand, represents a statistics statement indicating the presence of an effect or a difference”. The null and alternative hypotheses are exclusive; it is the alternative hypothesis that is tested and any conclusion regarding the null hypothesis is drawn by inference.
113 The classical model also requires a significance level to be chosen, although it should be noted Neyman & Pearson (1933) opposed the use of an arbitrary significance level, believing the researcher should choose the level in order to balance the risk of Type I and Type II errors. Fisher (1925) proposed 1% and
5%, although Fisher (1955) stated that it was not necessary to stipulate a significance level before an experiment (a modification of his earlier stance) and if the result was considered significant by the researcher, then the result should be reported along with its probability value (or p-value). Fisher’s concluding comment in that paper was: “We have the duty of formulating, of summarising, and of communicating our conclusions, in intelligible form, in recognition of the right of other free minds to utilize them in making their own decisions” seems to be a call for transparency as opposed to an arbitrary selection of a significance levels.
To avoid arbitrariness in this thesis, a probability level for the statistical test (‘p- level’) is calculated using both parametric and non-parametric tests and then interpreted. For guidance in interpretation, Bowerman et al. (2011: 360) provide the following interpretations of p-values: “0.1, some evidence; 0.05, strong evidence; 0.01, very strong evidence; 0.001, extremely strong evidence”, however the authors go onto note: “there are no really sharp borders between different weights of evidence. Rather, there is only increasingly strong evidence...as the p-value decreases”.
In the unusual cases here there is divergence in interpretation between the parametric and non-parametric tests, then each case is considered on its merits and a retrospective check has been made for consistency of interpretation across cases.
114 In summary the method proposed by Fisher (1955) and the terminology of
Bowerman et al. (2011) has been drawn upon to develop an approach to the interpretation of both the parametric and non-parametric results.
7.1.2 Set 1, Hypothesis 1: Returns to Scale
Hypothesis 1 is to establish whether pharmaceutical R&D demonstrates CRS or VRS. This allows for the selection of the most appropriate form of DEA model to be used in the subsequent hypothesis tests and is a necessary step in this research because previous literature has provided disparate results, with a sector-specific reference: Graves & Langowitz (1993) indicating
Decreasing Returns to Scale (DRS) in contrast to the generic Schumpeterian hypothesis of Increasing Returns to Scale (IRS) for R&D, as originally proposed in Schumpeter (1950).
Given the conflicting prior references on DRS and IRS, the null hypothesis does not presume either and is:
H1n: There are CRS for pharmaceutical R&D.
The alternative hypothesis, which will be tested, is:
H1a: There are VRS for pharmaceutical R&D.
A testing approach has been developed that avoids the statistical testing of
DEA scores by using the DEA scores to form two subgroups of the more efficient and less efficient, and testing the difference of the means of the R&D expenditure of the two subgroups.
The testing approach does presume that IRS and DRS do not exist simultaneously for different sizes of firms in the sample. However, examination
115 of the descriptive statistics in Figure 6.1 shows a decreasing monotonic relationship between scale efficiency and size of R&D expenditure (average).
7.1.3 Set 2, Hypotheses 2, 3 & 4: Firm Acquisition History and
Technical Efficiency
The main focus of the research is the association between acquisition history and the technical efficiency of the R&D process. The null hypothesis is that firms with above-median efficiency and below-median efficiency have similar merger history, as measured by the parameter NDV: the sum of M&A value over time divided by a normalisation factor that represents the size of the firm.
The set of firms is divided into two equal-sized subgroups of above- and below-average efficiency and the MCT, that is the mean or median3, of NDV for each subgroup is calculated. The null hypothesis is:
H2n: Firms with an above-median technical efficiency have the same
MCT of NDV as those with below-median technical efficiency.
The alternative hypothesis, which will be tested, is one-sided to reflect the
Merger Paradox and is:
H2a: Firms with an above-median technical efficiency have a lower
MCT of NDV than those with below-median technical efficiency.
Less formally, this states that companies that are more merger-prone are less efficient, as is consistent with the Merger Paradox.
3 The term MCT is used in preference to either mean or median because both parametric and non- parametric tests have been used to test the differences in MCT between samples, with the former testing mean and the latter testing median.
116 Two further hypotheses examined diversifications to gain resources in new markets (i.e. cross-border acquisitions) and new sectors (i.e. cross-sector acquisitions). The two hypotheses consider each of these types of diversification and follow a similar format to that considering acquisitions in general. The two null hypotheses are:
H3n: Firms with an above-median technical efficiency have the same
MCT of NDV for cross-border deals as those with below-median
technical efficiency.
H4n: Firms with an above-median technical efficiency have the same
MCT of NDV for cross-sector deals as those with below-median
technical efficiency.
The corresponding alternative hypotheses are:
H3a: Firms with an above-median technical efficiency have a lower
MCT of NDV for cross-border deals than those with below-median
technical efficiency.
H4a: Firms with an above-median technical efficiency have a lower
MCT of NDV for cross-sector deals than those with below-median
technical efficiency.
The outcomes of the testing of these alternative hypotheses are then compared with the outcomes of the tests examining financial efficiency, as described below.
7.1.4 Set 3, Hypotheses 5, 6 & 7: Deal History and Financial Efficiency
The null hypotheses H5n, H6n and H7n are restatements of H2n, H3n and
H4n but with ‘technical efficiency’ replaced with ‘financial efficiency’. However
117 the alternative hypotheses are phrased in the opposite direction, namely that above-median financial efficiency is associated with a higher MCT of NDV, as might be expected if the M&A deal is allowed to proceed and indeed Higgins
& Rodriguez (2006) confirm improved financial performance following acquisitions in the pharmaceutical industry.
The alternative hypotheses are:
H5a: Firms with an above-median financial efficiency have a higher
MCT of NDV than those with below-median financial efficiency.
H6a: Firms with an above-median financial efficiency have a higher
MCT of NDV for cross-border deals than those with below-median
technical efficiency.
H7a: Firms with an above-median financial efficiency have a higher
MCT of NDV for cross-sector deals than those with below-median
technical efficiency.
Financial efficiency is measured by ROS and ROA so six tests are undertaken on the three alternative hypotheses.
7.1.5 Set 4, Hypotheses 8, 9 & 10: Acquisition History and Sectoral
Efficiency
Hypotheses H8n, H9n and H10n are restatements of H2n, H3n and H4n but with NDV replaced by Sum of Deal Value (SDV). SDV omits the normalisation used to express the sum of previous acquisitions in a form relative to the size of the firm. Efficiency is still measured at firm level at the firm but all the acquisitions in the sector are associated with firms grouped into the more efficient and the less efficient. In practice the larger acquisitions made by the
118 larger firms overshadow the smaller acquisitions made by the smaller firms.
The effect is to examine if the sector’s acquisitions as a whole are associated with higher or lower efficiency at firm level and hence consider the effect of acquisitions on the sector in aggregate.
The direction of the alternative hypotheses reflects findings in the financial
M&A literature that M&A could benefit the sector as a whole, if not the acquiring firm; that is, the hypotheses are one-sided in the direction indicated by Seth et al. (2000) who found M&A provides benefits if the gains of the acquirer and acquired are considered together
The null hypotheses are that there is no difference in distribution of acquisition value between the more efficient and the less efficient companies. The alternative hypotheses, which will be tested, are:
H8a: Firms with an above-median technical efficiency have a higher
MCT of SDV than those with below-median technical efficiency.
H9a: Firms with an above-median technical efficiency have a higher
MCT of SDV for cross-border deals than those with below-median
technical efficiency.
H10a: Firms with an above-median technical efficiency have a higher
MCT of SDV for cross-sector deals than those with below-median
technical efficiency.
The outcomes of the testing of these alternative hypotheses are then used to examine the Merger Paradox at sector-level.
119 7.1.6 Set 5, Hypotheses 11, 12, 13 & 14: Acquisition History and Sales
over Assets
The purpose of H11 is to examine the effect of M&A on financial metrics. The null hypothesis for M&A is that there is no difference in the acquisition history of firms between those with above-median SOA and those with below-median
SOA (there is no reason to expect a difference to occur) and the hypothesis for M&A in aggregate is:
H11n: Firms with an above-median SOA have the same MCT of NDV
as those with below-median SOA.
The null hypotheses for cross-border and cross-sector deals, H12 and H13 respectively, are similar.
The alternative hypotheses, which will be tested, are one-sided in the direction indicated by Boekestein (2009), and are:
H11a: Firms with an above-median SOA have a lower MCT of NDV
than those with below-median SOA.
H12a: Firms with an above-median SOA have a lower MCT of NDV for
cross-border deals than those with below-median SOA.
H13a: Firms with an above-median SOA have a lower MCT of NDV for
cross-sector deals than those with below-median SOA.
This presumes that M&A leads to a greater recognition of intangible assets and hence a lowering of SOA, as Boekestein (2009) noted.
It transpires that in the last case the alternative hypothesis is ‘accepted’ but in the reverse direction, which undermines the basis for a unidirectional test.
120 Therefore a fourteenth non-directional alternative hypothesis has been formulated:
H14a: Firms with an above-median SOA do not have the same MCT of
NDV for cross-sector deals as those with below-median SOA.
H14n is identical to H13n, which is similar to H11n, as shown above.
7.2 Returns to Scale (H1)
The relationship between scale efficiency and an R&D scale variable has already been examined graphically and a monotonic relationship observed between the log of the R&D scale variable and the log of the scale efficiency, indicating VRS (more specifically, DRS). The hypothesis of CRS is now formally tested.
A two-sided parametric Welch test and a two-sided non-parametric Mann–
Whitney test for H1a was undertaken and in both cases H1a was accepted for p < 0.1%. Because the alternative hypothesis was accepted, the null hypothesis of CRS was rejected with strong statistical evidence.
The testing results are summarised in Table 7.1
Table 7.1 Results of Tests on Hypothesis H1a
Test P-value Interpretation
Non-parametric < 0.1% Extremely Strong Evidence to Accept
Parametric < 0.1% Extremely Strong Evidence to Accept
In summary there is extremely strong evidence for accepting VRS and the descriptive statistics confirm show this to be monotonic DRS and the null hypothesis H1n is rejected.
121 7.3 M&A and Technical Efficiency (H2, H3, H4)
7.3.1 Hypothesis 2
For the base model, the mean of the NDV for the companies with above- median efficiency is 0.91 and the mean NDV for the companies with below- median efficiency is 0.74, the reverse of that indicated in the alternative hypothesis. Testing for the significance of the reverse of H2a using a one- sided Mann–Whitney test gives a p-value of 28%. A one-sided Welch test gives a p-value of 30%. We conclude that there is no statistical evidence to accept the reverse of H2a on the basis of the DEA scores of the base model.
For the model with staff as an input, the mean of the NDV for the companies with above-median efficiency is 0.92 and the mean NDV for the companies with below-median efficiency is 0.72, a slightly larger difference than the base model, and the reverse of that indicated in the alternative hypothesis. Testing for the significance for the reverse of H2a using a one-sided Mann–Whitney test gives a p-value of 24%. A one-sided Welch test gives a p-value of 28%.
We conclude that there is no statistical evidence to accept the reverse of H2a on the basis of the DEA scores of the DEA model with staff inputs.
The testing results for H2a are summarised in Table 7.2:
Table 7.2 Results of Tests on Hypothesis H2a
Test Model Direction P-value Interpretation
Non-parametric, H2a Base Reverse 28% No Evidence
Parametric, H2a Base Reverse 30% No Evidence
Non-parametric, H2a Staff Input Reverse 24% No Evidence
Parametric, H2a Staff Input Reverse 28% No Evidence
122 In summary, in each case there is no statistical evidence to accept the reverse of the alternative hypothesis with either DEA model and the null hypothesis
H2n therefore stands.
7.3.2 Hypothesis 3
For the base model, the mean of the NDV for the companies with above- median efficiency is 0.21 and the mean NDV for the companies with below- median efficiency is 0.33, in the direction of that indicated in the alternative hypothesis. Testing for significance using a Mann–Whitney test gives a p- value of 40%. A one-sided Welch test gives a p-value of 28%. We conclude that there is no statistical evidence to accept H3a on the basis of the DEA scores of the base model.
For the model with staff as an input, the mean of the NDV for the companies with above-median efficiency is 0.18 and the mean NDV for the companies with below-median efficiency is 0.36, in the direction of that indicated in the alternative hypothesis. Testing for significance using a Mann–Whitney test gives a p-value of 47%. A one-sided Welch test gives a p-value of 20%. We conclude there is no statistical evidence to accept H3a on the basis of either test and therefore the null hypothesis stands.
The testing results for H3a are summarised in Table 7.3:
123 Table 7.3 Results of Tests on Hypothesis H3a
Test Model Direction P-value Interpretation
Non-parametric, H3a Base Standard 40% No Evidence
Parametric, H3a Base Standard 28% No Evidence
Non-parametric, H3a Staff Input Standard 47% No Evidence
Parametric, H3a Staff Input Standard 20% No Evidence
In summary, in each case there is no statistical evidence to accept the alternative hypothesis with either DEA model and the null hypothesis H3n therefore stands.
7.3.3 Hypothesis 4
For the base model, the mean of the NDV for the companies with above- median efficiency is 0.03 and the mean NDV for the companies with below- median efficiency is 0.06, in the direction of that indicated in the alternative hypothesis. Testing for significance using a Mann–Whitney test gives a p- value of 30%. A one-sided Welch test gives a p-value of 11%. We conclude, although there is formally no evidence on the basis of the parametric test
(although the p-value is close to the 10% threshold), and no evidence from the non-parametric test, taking the tests together, there is not sufficient statistical evidence to accept the alternative hypothesis. The null hypothesis therefore stands.
For the model with staff as an input, the mean of the NDV for the companies with above-median efficiency is 0.03 and the mean NDV for the companies with lower-median efficiency is 0.06, in the direction of that indicated in the alternative hypothesis. Testing for significance using a Mann–Whitney test gives a p-value of 31%. A one-sided Welch gives a p-value of 12%. We
124 conclude that there is not sufficient statistical evidence to accept the alternative hypothesis. The null hypothesis therefore stands.
The testing results for H4a are summarised in Table 7.4:
Table 7.4 Results of Tests on Hypothesis H4a
Test Model Direction P-value Interpretation
Non-parametric, H4a Base Standard 30% No Evidence
Parametric, H4a Base Standard 11% No Evidence
Non-parametric, H4a Staff Input Standard 31% No Evidence
Parametric, H4a Staff Input Standard 12% No Evidence
In summary, in each case there is no statistical evidence to accept the alternative hypothesis with either DEA model and the null hypothesis H4n therefore stands.
7.3.4 Summary of the Set
The main finding is that the DEA efficiency scores, for either DEA model, do not provide statistically significant evidence to accept the alternative hypothesis (or, where appropriate its reverse). Therefore the null hypothesis stands for all the three cases, namely M&A deals in aggregate, or cross- border and cross-sector deals, specifically are not associated with changes in technical efficiency.
Furthermore, the p-values for testing with the R&D-only model and the model with staff inputs are similar, and this suggests that any staff reduction effect is small. Therefore the base model alone is used in later hypothesis testing.
125 7.4 M&A and Financial Efficiency (H5, H6, H7)
7.4.1 Hypothesis 5
When financial efficiency is measured with ROS, the mean of the NDV for the companies with above-median efficiency is 1.15 and the mean NDV for the companies with below-median efficiency is 0.5, in the direction indicated by the alternative hypothesis. Testing for significance using a Mann–Whitney test gives a p-value of 20%. However a one-sided Welch test gives a p-value of
3%, which indicates strong evidence. Given the disparity between strong and no evidence, the interpretation of the results is that there is some evidence
(we note the average of the scores in the region of 10%, the threshold for
‘some evidence’).
When financial efficiency is measured with ROA, the mean of the NDV for the companies with above-median efficiency is 0.95 and the mean NDV for the companies with below-median efficiency is 0.7, again in the direction indicated in the alternative hypothesis. Testing for significance using a Mann–Whitney test gives a p-value of 33%. A one-sided Welch test on the NDV gives a p- value of 23%. We conclude that there is no statistical evidence to accept H5a for ROA and the null hypothesis stands.
The testing results for H5a are summarised in Table 7.5:
126 Table 7.5 Results of Tests on Hypothesis H5a
Test Model Direction P-value Interpretation
Non-parametric, H5a ROS Standard 20% No Evidence
Parametric, H5a ROS Standard 3% Strong Evidence
Non-parametric, H5a ROA Standard 33% No Evidence
Parametric, H5a ROA Standard 23% No Evidence
In summary, following the discussion above, there is some statistical evidence to accept the alternative hypothesis when ROS is used but not when ROA is used to measure financial efficiency. The null hypothesis H5n is therefore rejected for ROS but stands for ROA.
7.4.2 Hypothesis 6
When financial efficiency is measured with ROS, the mean of the cross-border
NDV for the companies with above-median efficiency is 0.41 and the mean
NDV for the companies with below-median efficiency is 0.13, in the direction of that indicated in the alternative hypothesis. Testing for significance using a
Mann–Whitney test gives a p-value of 12% for acceptance of the alternative hypothesis. A one-sided Welch test gives a p-value of 8%. Taking both results together, the parametric result indicates some evidence although the non- parametric does not; however the non-parametric result is only slightly over the threshold and the mean of the results is on the threshold. On this basis, there is some statistical evidence to accept H6a and on the basis of the ROS scores and reject the null hypothesis.
When financial efficiency is measured with ROA, the mean of the cross-border
NDV for the companies with above-median efficiency is 0.26 and the mean
NDV for the companies with below-median efficiency is 0.28, the reverse of
127 that indicated in H6a. Testing the reverse of H6A for significance using a
Mann–Whitney test gives a p-value of 22%. A one-sided Welch test gives a p- value of 47%. We conclude that for ROA there is no statistical evidence to accept the reverse of H6a on the basis of either a parametric or a non- parametric test.
The testing results for H2a are summarised in Table 7.6.
Table 7.6 Results of Tests on Hypothesis H6a
Test Model Direction P-value Interpretation
Non-parametric, H6a ROS Standard 12% No Evidence
Parametric, H6a ROS Standard 8% Some Evidence
Non-parametric, H6a ROA Reverse 22% No Evidence
Parametric, H6a ROA Reverse 47% No Evidence
In summary, there is some statistical evidence to accept the alternative hypothesis H6a when ROS is used but not to accept the reverse of H6a when
ROA is used to measure financial efficiency. Therefore where ROS is used the null hypothesis H6n is rejected but it stands when ROA is used.
7.4.3 Hypothesis 7
When financial efficiency is measured with ROS, the mean of the cross-sector
NDV for the companies with above-median efficiency is 0.02 and the mean
NDV for the companies with below-median efficiency is 0.06, the reverse of the direction indicated in the alternative hypothesis. Testing the reverse of H7a using a Mann–Whitney test gives a p-value of 33% and a one-sided Welch test gives a p-value of 8%. The results do diverge, however the non- parametric test is close to the 10% threshold and the mean of the results
128 exceeds the threshold by a factor of two. Given this there is no reliable statistical evidence to accept the reverse of hypothesis H7a.
When financial efficiency is measured with ROA, the mean of the cross-sector
NDV for the companies with above-median efficiency is 0.044 and the mean
NDV for the companies with below-median efficiency is 0.041: close to equal but the reverse of that indicated in H7a. A one-sided Welch test of the reverse of H7a gives a p-value of 45% and a Mann–Whitney test (which compares medians) gives a p-value of 18%. We note there is no statistical evidence to accept the reverse hypothesis H7a with either test.
The testing results for H2a are summarised in Table 7.7:
Table 7.7 Results of Tests on Hypothesis H7a
Test Model Direction P-value Interpretation
Non-parametric, H7a ROS Reverse 33% No Evidence
Parametric, H7a ROS Reverse 8% No Evidence
Non-parametric, H7a ROA Reverse 18% No Evidence
Parametric, H7a ROA Reverse 45% No Evidence
In summary, there is no reliable evidence to accept the reverse of alternative hypothesis H7a for either ROA or ROS when the parametric and non- paramagnetic tests are considered in unison. The null hypothesis H7n therefore stands.
7.4.4 Summary of the Set
For M&A in aggregate and for cross-border deals there is some statistical evidence to accept the alternative hypotheses H5a and H6a, and hence reject the hypotheses H5n and H6n for M&A in aggregate and cross-border deals
129 respectively, when ROS is used as a metric for financial efficiency; however this is not observed when ROA is used. For cross-sector deals, there is no evidence to accept H7a with either ROS or ROA.
7.5 Sector Effects and Technical Efficiency (H8, H9, H10)
7.5.1 Hypothesis 8
For the base model, the mean of the SDV for the companies with above- median efficiency is 19174 and the mean SDV for the companies with below- median efficiency is 4345, in the direction of that indicated in the alternative hypothesis. Testing for significance using a Mann–Whitney test gives a p- value of 18%. A one-sided Welch test gives a p-value of 4%. Following similar logic to that proposed for the testing of H5a for the use of ROS, namely the parametric test shows strong evidence although the mean of the parametric and non-parametric tests is in the region of 10%, we conclude that there is some statistical evidence to accept H8a on the basis of the DEA scores of the base model. The testing results for H8a are summarised in Table 7.8:
Table 7.8 Results of Tests on Hypothesis H8a
Test Direction P-value Interpretation
Non-parametric, H8a Standard 18% No Evidence
Parametric, H8a Standard 4% Strong Evidence
On this basis H8n was rejected.
7.5.2 Hypothesis 9
For the base model, the mean of the SDV for the companies with above- median efficiency is 4442 and the mean NDV for the companies with below-
130 median efficiency is 1056, in the direction of that indicated in the alternative hypothesis. Testing for significance using a Mann–Whitney test gives a p- value of 26%. A one-sided Welch test gives a p-value of 4%. As the mean of these scores is 15%, we conclude that there is no statistical evidence to accept H9a or reject the null hypothesis.
Table 7.9 Results of Tests on Hypothesis H9a
Test Direction P-value Interpretation
Non-parametric, H9a Standard 26% No Evidence
Parametric, H9a Standard 4% Strong Evidence
There is no statistical evidence to accept the hypothesis H9a and the null hypothesis H9n stands.
7.5.3 Hypothesis 10
For the base model, the mean of the SDV for the companies with above- median efficiency is 744 and the mean SDV for the companies with below- median efficiency is 472, in the direction of that indicated in the alternative hypothesis. Testing for significance using a Mann–Whitney test gives a p- value of 45%. A one-sided Welch test gives a p-value of 30%. We conclude that there is no statistical evidence to accept H10a. The testing results for
H10a are summarised in Table 7.10:
Table 7.10 Results of Tests on Hypothesis H10a
Test Direction P-value Interpretation
Non-parametric, H10a Reverse 28% No Evidence
Parametric, H10a Reverse 30% No Evidence
131 There is no statistical evidence to accept the hypothesis H10a and the null hypothesis H10n stands.
7.5.4 Summary of the set
The main finding is that in aggregate the DEA efficiency scores do provide some statistically significant evidence to accept the alternative hypothesis that mergers are associated with higher technical efficiency in aggregate, when
SDV was considered to examine sector effects.
For cross-border deals the parametric test gave strong evidence to accept but this was not supported by the non-parametric test and therefore a conclusion of no evidence was drawn; for cross-sector deals neither test gave evidence to accept the alternative hypotheses.
7.6 M&A and Financial Metrics (H11, H12, H13, H14)
7.6.1 Hypothesis 11
For M&A in aggregate, the mean SOA of the cross-sector NDV for the companies with above-median efficiency is 0.62 and the mean NDV for the companies with below-median efficiency is 1.03, in the direction of the alternative hypothesis. Testing for significance using a Mann–Whitney test gives a p-value of 40%. A one-sided Welch test gives a p-value of 12%.
The testing results for H11a are summarised in Table 7.11:
Table 7.11 Results of Tests on Hypothesis H11a
Test Direction P-value Interpretation
Non-parametric, H4a Standard 40% No Evidence
Parametric, H4a Standard 12% No Evidence
132 There is no statistical evidence to accept the hypothesis H11a and the null hypothesis H11n stands.
7.6.2 Hypothesis 12
For cross-border deals, the mean SOA of the cross-border NDV for the companies with above-median efficiency is 0.23 and the mean NDV for the companies with below-median efficiency is 0.31, in the direction of the alternative hypothesis. Testing for significance using a Mann–Whitney test gives a p-value of 45%. A one-sided Welch test gives a p-value of 35%. The testing results for H12a are summarised in Table 7.12:
Table 7.12 Results of Tests on Hypothesis H12a
Test Direction P-value Interpretation
Non-parametric, H12a Standard 45% No Evidence
Parametric, H12a Standard 35% No Evidence
There is no statistical evidence to accept H12a and the null hypothesis H12n stands.
7.6.3 Hypotheses 13 & 14
For cross-sector deals, the mean SOA of the cross-sector NDV for the companies with above-median efficiency is 0.07 and the mean NDV for the companies with below-median efficiency is 0.02, in the reverse direction of the alternative hypothesis. If the reverse of H13a is tested, namely “For cross- sector deals, firms with an above-median SOA have a higher MCT of NDV than those with below-median SOA”, the significance using a Mann–Whitney test gives a p-value of 7%. A one-sided Welch test gives a p-value of 3%.
Taken together, there is strong evidence to accept the reverse of hypothesis
133 H13a because the parametric test indicates strong evidence and the mean of the p-values is below the 5% threshold. The testing results for H13a are summarised in Table 7.13:
Table 7.13 Results of Tests on Hypothesis H13a
Test Direction P-value Interpretation
Non-parametric, H4a Reverse 7% Some Evidence
Parametric, H4a Reverse 3% Strong Evidence
To summarise the reverse of the alternative hypothesis H13a has been accepted; however Sheshkin (2011: 451) notes this cannot be a basis for rejecting a null hypothesis. A non-directional alternative hypothesis, which will be tested instead, is therefore tested.
Testing for significance using a Mann–Whitney test gives a p-value of 14%. A one-sided Welch test gives a p-value of 6%. As the mean of these scores is
10%, we conclude that there is some (but not strong) statistical evidence to accept H14a. The results are shown in Table 7.14:
Table 7.14 Results of Tests on Hypothesis H14a
Test P-value Interpretation
Non-parametric, H14a 14% No Evidence
Parametric, H14a 6% Strong Evidence
The alternative hypothesis H14a is accepted and the hull hypothesis H14n is therefore rejected. Because H13n is the same as H14n this is also rejected.
134 7.6.4 Summary of the Set
For H11 and H12, although the ratio of the means is in the direction that would accept the alternative hypothesis and hence reject the null hypotheses, there is insufficient statistical evidence to do so.
However, the cross-sector finding for H13 does show a statically significant finding, associating increased SOA with increased cross-sector activity, in the reverse direction of H13a. The null hypothesis cannot be rejected on this basis and so the non-directional equivalent, H14a, was tested instead and some evidence was found to accept this hypothesis and reject the null hypothesis.
The implications for financial measure selection are discussed in the next chapter.
7.7 Summary of Findings
7.7.1 Collation of Findings
Table 7.15 summarises those tests where there was evidence to reject the null hypothesis.
Table 7.15 Rejected Null Hypotheses
H1n There are CRS for pharmaceutical R&D
H5n Firms with an above-median financial efficiency have the same
MCT of NDV as those with below-median financial efficiency.
H6n Firms with an above-median financial efficiency have the same
MCT of NDV for cross-border deals as those with below-median
financial efficiency.
135 H8n Firms with an above-median technical efficiency have the same
MCT of SDV as those with below-median technical efficiency.
H13n, For cross-sector deals, firms with an above-median SOA have the
H14n same MCT of NDV as those with below-median SOA.
All were rejected with some evidence, except for H1n which was rejected with extremely strong evidence.
7.7.2 Check for Consistency
Both parametric and non-parametric tests have been considered and in most cases they are in agreement on the presence or absence of statistical evidence. However there are cases where the p-value of the parametric test is below the threshold and the p-value of the non-parametric test is above the threshold and each case has been assessed on its merits. The treatment is however consistent and in all cases follows the rule:
If the p-value of the parametric test is below the threshold for evidence
and the mean of the parametric and non-parametric p-values is above the
threshold by no more than a fifth of the threshold value, then the test for
evidence is accepted.
The treatment of borderline cases is summarised below in Table 7.16.
136 Table 7.16 Borderline Alternative Hypotheses
No. Hypothesis Parametric Mean p-
p-value value
Some evidence
H5a Firms with an above-median financial 3 11.5
efficiency [ROS] have a higher MCT of NDV
than those with below-median financial
efficiency.
H6a Firms with an above-median financial 8 10
efficiency [ROS] have a higher MCT of NDV
for cross-border deals than those with below-
median technical efficiency. (When using
ROS)
H8a Firms with an above-median technical 4 11
efficiency have a higher MCT of SDV than
those with below-median technical efficiency.
H14a For cross-sector deals, firms with an above- 4 10
median SOA does not have the same MCT of
NDV as those with below-median SOA.^
No evidence
137 H7a Firms with an above-median financial 8 20.5
efficiency [ROS] have the same MCT of NDV
for cross-sector deals as those with below-
median financial efficiency.
H9a Firms with an above-median technical 4 15
efficiency have a higher MCT of SDV for
cross-border deals than those with below-
median technical efficiency.
^ H14a was assessed as ‘some evidence’ not ‘strong evidence’ because the mean p-value was more than one fifth above the 5% threshold for ‘strong evidence’.
Table 7.16 shows a consistent treatment of cases that seeks to balance the risk of Type I errors (concluding a false alternative hypothesis is true by a adopting a lax threshold) and a Type II error (concluding a true alternative hypothesis is false by adopting a strict threshold). It can be seen that the borderline has been drawn between the results for H8a and H9a: in both cases the parametric p-value was 4% but in the former the mean of the parametric and non-parametric tests was 11% judged to support a view of some evidence but in the latter the mean was 15%, judged to represent no evidence.
138 8 Discussion of Results
8.1 Range of Findings
The five groups of hypotheses have addressed five different aspects of the measurement of PAP:
– The returns to scale of the pharmaceutical R&D process. It is necessary
to establish this in order to select the most appropriate model for the
measurement of R&D comparative efficiency. This has yielded a very
strong conclusion on returns to scale for the process and shows DRS.
– The association of technical efficiency with M&A history. This has not
yielded statistically significant results, although this is itself a finding
(discussed later) and an indirect contribution to the understanding of the
Merger Paradox when contrasted with the findings on financial efficiency.
– The association of financial efficiency with M&A history. These findings
differ according to the financial measure chosen. However, a rationally
selected measure, ROS, does show a relation that is in opposition to the
merger paradox.
– The sector-level consequences of M&A. It has been found that
substantially more acquisitions by value are associated with technically
efficient firms than with less efficient firms.
– The examination of the interrelationship between two measures of
financial efficiency for the different types of deal. The results of the test
shows the choice of financial metric for PAP can lead to differing
conclusions and the ratio of ROS to ROA, SOA, was affected by the level
of historical cross-sector acquisitions.
139 These findings on measures themselves depend on a robust process for measure selection and are therefore supported by the systematic approach to the selection of measures encapsulated in the 12 Design Principles.
Each of the five areas of results is now discussed in turn, focusing on the findings, followed by a section on the qualitative findings for the selection of measures for PAP (the Design Principles themselves were designed for a general application, not specifically PAP).
The contributions arising from the findings are discussed in the following section.
8.2 Returns to Scale of R&D
The strength of the linear relationship, when R&D output is measured by the number of compounds, between the logarithm of the scale efficiency of R&D and the logarithm of R&D expenditure (i.e. a power law) was a striking feature, when perhaps more variability might have been expected given the role of serendipity in R&D. VRS were established with p < 0.01% and the clarity of the power-law relationship between returns to scale and scale efficiency may have future econometric application.
The relationship was less clear when the numbers of compounds were considered. A multiplicity of clinical trials, especially in the later stages when expense can be an issue, may not be an indicator of higher efficiency. The analysis of the ratio of clinical trials and compounds between companies in the later stages of the pipeline does suggest a divergence of management practice, whereas in the early stages when the safety of a compound has to be established, divergence is less and there are lower ratios of compounds to trials. The findings suggest (although in the absence of statistical tests) that
140 the number clinical trials should not be considered an output measure in preference to the number of compounds.
The DRS for pharmaceutical R&D has significant implications for the Merger
Paradox and these are discussed in the next sections.
8.3 The Association of Technical Efficiency with M&A History
The absence of a statistically significant relationship between technical efficiency and merger history might at first sight seem disappointing, although it is a finding in itself. There are many possible reasons for the absence of a relationship at the firm level, given the multiplicity of motives for M&A. Some acquirers choose to use the strength of their R&D pipeline, which is reflected in market ratings and hence share price, to acquire potential competitors; in these cases, there would be a strong association between acquisition history and R&D efficiency. Conversely, there are other cases, where historical M&A deals have been undertaken with the objective of achieving economies of scale by the reduction of overheads costs in the face of weak pipelines; in these cases an inverse relationship might be expected with M&A history and technical efficiency. Examples of both were given in Section 6.12: Pfizer is an example of an aggressive acquirer and the Sanofi-Aventis merger is an example of two low productivity firms merging, with the lowering of fixed costs being a plausible motive.
The methodology of this study has not included an event study so cannot comment directly on whether M&A is damaging to the acquiring company.
Nonetheless the findings on economies of scale are unequivocal and because an M&A deal inevitably leads to a larger company it can be expected that R&D productivity will fall. Because the future prosperity of a pharmaceutical company depends on its R&D productivity compared with competitors, it can
141 be stated indirectly that the Merger Paradox has been confirmed in the pharmaceutical sector: pharmaceutical M&A deals are popular but do not improve the performance of the acquirer.
Similar comments also apply to cross-border mergers, tested by H3, and cross-sector deals, tested by H4.
8.4 The Association of Financial Efficiency with M&A History
ROS was selected over ROA as a better measure of financial efficiency; this was done because of the distortions, noted in the literature, inherent in ROA when it is used to measure financial efficiency in a sector with substantial intangible assets.
A statistically significant relationship has been established between financial efficiency of the firm as measured by ROS, and that an improved ROS is associated with higher historic merger activity; this is in contrast to the findings for technical efficiency where no statistically significant relationship was found.
The contrast is further support of the Merger Paradox because although the acquirer may have many motives for the deal (including an increase in ROS), the change in financial efficiency is also likely to act as a qualifying factor, for instance a deal that lowers financial efficiency is unlikely to proceed even if there are other benefits. We therefore observe a divergence in incentives for
M&A, with an association with an increase in an accounting measure (ROS) but no equivalent prospect of an increase in long-term operational performance (technical R&D efficiency).
This argument is an elaboration of Angwin (2007) in which the multiplicity of motives was recognised; although such a multiplicity undoubtedly exists and even if the motives are not metric-related (e.g. the motive of elimination of a
142 competitor to enhance market power), metrics can also act as a qualifying factor for a deal to proceed. It is therefore necessary to consider a two-phase model of M&A comprising initial motives and subsequent metric-related constraints or qualifying factors, rather than motive alone.
There were very similar findings for cross-border deals; this is perhaps not surprising because international firms now dominate the pharmaceutical industry and the importance of the ‘cross-border’ effect may be attenuated and many major cross-border deals did follow language-orientations, for example
USA/UK deals.
However, the ROS for cross-sector deals had no relation to acquisition history and we later suggest that this arises because of the difference in ROS between the pharmaceutical sector and the sectors in which pharmaceutical companies make acquisitions.
Regarding the lack of statistically significant findings for ROA, this can be explained by reference to the unreliability of the denominator, where investments in intangible assets are not recognised as assets when the assets are internally created. Even if the numerator of the measure were improved for acquiring firms, then the effect of increased acquisitions leading to greater recognition of assets would depress the effect of an improved ratio as reported in the literature. The effect of acquisitions on the two measures is tested directly later, where this effect has been observed but not to a level of statistical significance.
8.5 The Sectoral Consequences of M&A
This thesis does not examine sector effects directly but it is possible to associate the total value of deals with the surviving firms and their relative
143 efficiency. H8a was confirmed, namely ‘Firms with an above-median technical efficiency have a higher MCT of SDV than those with below-median technical efficiency.’ This does have implications for the sector because it is has been established that more acquisitions by monetary value are associated with efficient firms than with the less efficient firms. The implication is that those large firms who choose to acquire tend to be more efficient than small acquiring firms and non-acquirers.
At the sectoral level, the implication is that M&A activity, where it does occur, leads to a concentration of market power in the more efficient firms, which is the underlying argument for a liberal M&A regime. There is a parallel with findings from the financial M&A literature, where M&A is found to increase the total wealth of shareholders in the acquiring and acquired firm, even if the subsequent performance of the acquirer is indifferent.
For the individual pharmaceutical firm, however, it can be expected to become less technically efficient following the acquisition because R&D efficiency decreases with size, and if this occurs it may itself be acquired in the future.
8.6 Relationship between SOA and M&A
The testing of SOA for M&A in aggregate and for cross-border deals did not lead to a higher NDV for firms with below-median SOA being found to be statistically significant, even though this might have been expected from consideration of accounting principles and previous academic literature.
For cross-sector M&A, however, there was strong statistical evidence that firms with above-median SOA had a higher cross-sector NDV.
This is explicable by reference to industrial practice. Some sector acquisitions tend to be into sectors that have operations elsewhere in the health value
144 chain, especially companies that trade in medical goods, in order to obtain a route to market. These companies tend to have elevated SOA ratios because they are distributors and retailers, and exhibit a high turnover of goods through the supply chain with a relatively small asset base (the retail and distribution network).
8.7 Synthesis of Findings
The previous discussion has considered the implications of each quantitative finding individually and some are valuable in their own right, for example the establishment of a linear relationship between the logarithm of scale efficiency and the logarithm of R&D expenditure. We now synthesise the findings in order to establish their relevance to the main topic of this thesis: the measurement of PAP in the pharmaceutical sector, as it relates to R&D.
The scale efficiency finding is important on several fronts. Firstly, it shows that
M&A, which leads to larger companies, can be expected to lower efficiency unless off-setting gains in technical efficiency are to be found; no statistically significant evidence has been found that this is the case. This finding is therefore supportive of the Merger Paradox, namely an activity is continuing which can be expected to lead to a lowering of performance in a crucial business process.
However, the finding also provides a motive for the continuation of the practice. If a large firm recognises that its own R&D is unproductive it is in the position to temporarily redress this in the short term by the purchase of smaller productive companies, even though both the acquiring and the acquired company might have concerns regarding the effect on longer-term performance. This is especially problematical in R&D where the resources being purchased are largely intangible and therefore can be readily dissipated
145 following the merger, for example by the resignation of key staff. We therefore see an additional pharmaceutical specific Merger Paradox, namely the response of a pharmaceutical firm to declining R&D is likely to exacerbate the problem. This also has implications for public policy, especially because the productivity of R&D is declining at the sectoral level.
We now turn to the finding that if M&A in the sector as a whole is considered, then it is associated with the most efficient companies; this may seem to suggest that at the sectoral level M&A is functioning as it is intended, namely it leads to a concentration of power with the more successful firms and the elimination of the poorer performers. That may well be the case, because it is possible that the share prices of the efficient firms permit them to make hostile takeovers to increase their market power and eliminate competition, while
M&A permits the less productive to undertake non-hostile deals with a shared objective of reducing fixed costs to improve efficiency. Therefore at a given moment in time, the observation that the sector’s acquisitions are associated with more technically efficient companies is not inconsistent with the finding that historically M&A tends to lower performance. A similar effect is observed at the disaggregated product-level within a pharmaceutical firm, where the most successful products or technologies today are unlikely to remain so as patents expire and new competition emerges; nonetheless despite the certainty of this occurring it may still be rational to focus R&D investments on the areas that are currently most successful in order to maximise the return on
R&D in the short term.
Although the motives for an M&A deal are diverse, the deals are unlikely to proceed if they damage the financial performance of the firm. The findings that acquisitions are associated with firms that are more financially efficient is therefore fully consistent with the Merger Paradox because improved financial
146 performance will act as a qualifying factor for a deal proceeding, whatever its subsequent effect on the underlying fundamentals of the firm.
In summary, the findings point not so much to a Merger Paradox as to the
Paradoxes of Mergers. Not only do M&A deals continue when their effects on long-term performance are likely to disappoint (if only because of scale effects) but the behaviour seems fully rational to the acquirer; furthermore the need for good accounting-based performance will ensure that when analysed by conventional accounting measures, M&A will seem to be successful, at least in the short term. Meanwhile, at the sector-level, it would seem that the most efficient companies undertake acquisitions at any moment in time, even though over time they may become less efficient as a result.
Moving on from the role of measurement in the Paradoxes of Mergers and their motives to the narrower topic of measurement of PAP itself, the findings highlight the difficulties of PAP measurement, especially where intangibles are involved. The effect on M&A of the recognition of intangibles in ROA was noted in earlier literature and was not refuted in this thesis. However, a separate effect on the preferred alternative measure, namely ROS, has been identified when cross-sector deals are considered, this effect arises from diversification into retail and distribution from research-based manufacturing.
This reinforces the need for a PMF when studying M&A rather than relying upon a single measure. Regarding financial measures, neither ROA nor ROS should be relied upon on their own and non-financial measures are required to consider the preservation of intangibles in the aftermath of a deal.
Furthermore, if a PMF for the external evaluation of M&A is to be used, then there should be a theoretical basis for the selection of the measures, and the
RBV-based 12 Design Principles have shown themselves to be a practical
147 approach through their application in designing a PMF for a pharmaceutical firm.
148 9 Contributions
9.1 Overview
This thesis examines the association of efficiency of the R&D process in the major pharmaceutical firms with their history of M&A. Studies of PAP are numerous, and their results diverse, however this thesis starts from the premise that PAP is itself an intellectual construct based upon assumptions on motives for the deal and the perspectives of relevant stakeholders. Therefore prior to measuring PAP there has to be a thorough consideration of the measures to be used.
From this initial stance on the assessment of PAP, this thesis contributes to the field by first applying the RBV to the selection of multiple performance measures for a PMF; this approach is encapsulated as a set of 12 Design
Principles that can be applied in any sector. The thesis then sheds new light on the Merger Paradox, namely why M&A continue to be transacted when historically their results seem to be disappointing overall. The thesis contrasts
PAP as measured by a PMF with PAP as measured by a conventional financial measure: ROS. In essence, an association between above-median
ROS and increased acquisition activity was established, but the same relationship was not established when a non-financial PMF was used. This finding provides an incentive-related explanation for the Merger Paradox linked to differing indications from financial and non-financial measures.
The thesis also examines PAP of subsets of acquisitions, namely cross-border and cross-sector deals, to consider diversification effects and establishes a contrast between the findings as they affect the individual firms and the effects
149 at the sectoral level, which has parallels with earlier research using financial event studies.
By adopting a novel means for the assessment of PAP, namely combining a longitudinal view of acquisition history and a cross-sectional view of comparative efficiency (that itself considers the longitudinal nature of the R&D pipeline and the multiple outputs of the R&D process), this thesis has provided a new application for DEA in the M&A literature that has not previously been used to examine R&D. In doing so, the use of DEA has also established scale efficiency factors for pharmaceutical R&D, clarifying earlier ambiguity in findings in this field and recognising the multiple inputs and outputs of the process.
A further empirical contribution is the examination of the relation between size and frequency of M&A in the pharmaceutical sector and its consistency with a power-law distribution.
Finally, the thesis examines the relative merits of ROS and ROA as a financial measure and whether there is a statistically significant difference in the conclusions on PAP arising from the use of the two different measures.
9.2 Contributions to the Acquisition Literature
The timing of the findings of the research are fortuitous because after four decades of contradictory findings on PAP, there is now a focus on the assumptions underlying this construct and especially the examination of the original objectives for an acquisition (Angwin, 2007) and how one can measure attainment of those objectives. This is related to the Merger Paradox, which is concerned with the motives and their attainment, and currently the most common explanations involve an element of agency theory, namely the
150 divergence of motives between managers and shareholders. This research suggests that agency theory is not required to explain the Merger Paradox, at least in the pharmaceutical sector and the actions of managers are consistent with improving a common financial measure under their control.
9.2.1 Reduction of Uncertainty
In undertaking an additional study in a well-researched area where there is already a disparity of findings, there is a danger of adding another observation that does not provide further insight. This is especially the case where meta- analyses have concluded that previous studies, for example King et al. (2004), have not identified the full range of moderating factors on PAP, and others have concluded that there is little relation to the findings of studies where different measurement principles have been employed, for example
Schoenberg (2006) and Papadakis & Thanos (2010).
The line of enquiry in this research has therefore been to reduce uncertainty by consciously removing potential moderating factors from the research. By focusing on one industry and analysing performance on the same set of companies in two different ways, it was possible to establish that a multiparameter, non-financial method of measurement and a common financial measurement gave rise to differing conclusions. Although many other factors may affect performance, the conclusions on differences in performance as measured by financial and non-financial parameters have been established.
9.2.2 Differences in Measured Performance
The lack of a statistically significant association between technical efficiency and M&A history is a useful finding, especially when coupled with the
151 presence of statistically significant findings showing diseconomies of scale in
R&D and in associations between M&A and ROS.
Because M&A leads to larger entities with larger R&D processes, this might be expected to lead to a lower efficiency in the longer run, which might offset shorter-term cost reductions from the deal (e.g. removal of duplicated posts or premises). This is a natural explanation for the lack of association, coupled with the possibility that some companies with lower technical efficiency may choose to merge in order to seek short-term economies and buttress financial performance.
Regarding PAP as measured financially, companies with above-median ROS had a significantly higher historical incidence of acquisitions. A direct comparison with an event-study approach using financial measures cannot be made and indeed the findings of these event studies are not consistent amongst themselves, possibly because of unidentified moderating factors, as suggested by King et al. (2004).
A lack of association between ROA performance and M&A history may arise from the shortcomings of that measure where intangible assets are significant, despite it being the most popular accounting measure in PAP research. This is a worrying finding but nonetheless a contribution to the M&A literature.
The thesis sheds much light on the Merger Paradox, whereby acquisitions continue despite their disappointing non-financial outcomes (Schenk, 2008).
The transactions may improve ROS or be facilitated by higher ROS originally
(e.g. access to merger finance). This explanation is also consistent with
Schoenberg’s (2006) study in which it was suggested that approach to measurement was an explanation for the variation in the findings of research on PAP.
152 9.2.3 Diversification
The research also examined the impact of diversification and the findings here for cross-border deals were different from cross-border deals when ROS was considered.
For cross-border deals, there was an association of such deals with improvements in financial efficiency, similar to that for acquisitions as a whole, however the cross-sector analysis did yield a strongly significant result for
ROS. This is in line with Shelton (1988) who reported: “Multivariate regression analysis shows that acquisitions which permit the bidder access to new but related markets create the most value with the least variance” with cross- sector deals into retailing and wholesaling of health goods being associated with poorer performance, unlike cross-border deals which offer access to new markets for R&D-based products.
9.3 Contributions to the Performance Literature
There is a growing body of literature on performance measurement using multiple measures in PMFs for an individual firm, but the choice of measures has presumed a detailed knowledge of the internal operations of the company to select the parameters. The development of a theoretical basis for the solution of measures could assist in multiparameter measurement processes in general. It is especially useful in the field of comparative efficiency analysis that is frequently undertaken from outside the firm.
Any proposed theoretical approach should be capable of considering both tangible and intangible inputs and outputs and be related to a theoretical base that is widely accepted. The RBV meets both these requirements and has evolved into the dominant theory, however relatively little attention has been
153 paid to the measurement of resources within the RBV literature. By reviewing the literature within the RBV, where measures were considered, a set of criteria for selection of measures relevant to competitive advantage was identified and used as a set of Design Principles. As well as being used to select measures for this research, the Design Principles and Construction
Process could have further application in other sectors.
9.4 Contributions to the DEA Literature
The application of DEA to measure the effects of M&A has been confined to date to short-term event studies that are not suitable for measuring R&D efficiency in a sector that has a lengthy R&D pipeline. The analysis of the efficiency of the R&D pipeline using data on its inputs for the majority of its average duration and then using DEA in a cross-sectional analysis is a novel form of the use of DEA for the analysis of M&A. In other respects the use of
DEA is conventional although the application is unusual in the effort taken to measure intangible outputs directly rather than through the use of financial surrogates for intangibles, for example revenue or share price, although the pharmaceutical sector was consciously selected to enable this to be possible.
In summary, this research has extended the application of DEA, as well as considering its field of application and the approach taken to the selection of inputs and outputs to the DEA models.
9.5 Empirical Contributions
9.5.1 Returns to Scale
There have been earlier attempts in the literature to measure R&D productivity, for example Graves and Langowitz (1993) used a simple unit cost, examining approved compounds produced per unit of R&D expenditure,
154 but they did not account for the multiple outputs of the R&D process.
Therefore application of a multiple input/output approach to the R&D pipeline is itself a useful contribution.
DEA was selected even though it has not been extensively used to measure the efficiency of the R&D process of a firm. The use of DEA to measure the efficiency of the R&D process for pharmaceutical firms of a variety of sizes provided the opportunity to determine if pharmaceutical R&D had CRS. It was found not to and furthermore a clear relationship was found between scale efficiency and a R&D scale parameter, which showed DRS, in contrast to the
Schumpeterian hypothesis (over IRS in R&D) and Jensen (1987) who found that marginal productivity was not adversely affected by firm size.
From the view of industry practice, the industry has also been concerned with a fall in the number of new drugs approved despite rising R&D costs, although
Cockburn (2006) suggests this is due to rising R&D input costs rather than declining efficiency in conversion of inputs to outputs. The response of some companies to declining internal efficiency has been to acquire R&D resources externally, through acquisition, although this response is not universal.
Although acquisition will circumvent the effects of the R&D efficiency problem in the short term, this research suggests that it may only add to the problem of declining R&D efficiency in the long term, if the association with acquisitions and lower productivity reflects a causal link.
9.5.2 Power Laws in M&A
Further empirical findings included the distribution of mergers by size and establishing the limits of the power-law hypothesis, including the divergence of the mega-mergers from the linear log–log relationship for this particular sector
(an elaboration on Park et al., 2010). However the research has not
155 established a power law conclusively because it has not eliminated alternative explanations for the linear log–log relationship. The limits of the linear relationship in pharmaceutical M&A in the chosen sample involved a surfeit of mega-mergers and a lower than expected number of mergers in the size range immediately below the mega-merger range. Beneath these large sizes, a power law prevails indicating a self-organised critical system. This finding is consistent with industry analysis, where there has been a conscious attempt to consolidate the industry from the top, consisting of mergers of mid-tier pharmaceutical companies in the second M&A wave and then the merger of giants in the third M&A wave.
It was also found that following the removal of zeros from the statistics for
NDV, the statistical distribution follows a Poisson distribution, suggesting that the normalisation of total deal value by cost of sales reveals the underlying
Poisson process (and thus confirms the validity of the normalisation factor).
9.5.3 Financial Metrics in M&A
The divergence of findings between the financial and non-financial metrics and between different financial metrics was a major finding of the research and adds further support for the use of a PMF over a single measure
(Schoenberg, 2006). However there is also a case for the use of using multiple financial metrics, with the potential shortcomings of ROA being noted previously and the limitations of ROS for analysing cross-sector deals being established within this thesis. The finding relating to SOA, the ratio of ROS to
ROA, showed that particular types of deal (e.g. cross-sector deals) can have a differential effect on common financial metrics.
156 9.6 Directions of Future Research
Greater recognition of the role of chance in R&D productivity and the sensitivity of the findings to this random element would be a fruitful area for future research. The role of uncertainty in the analysis of DEA has been considered by Dyson & Shale (2010), and the use of Monte Carlo simulation to perturb the outputs from the R&D processes (the inputs are well defined) to observe the effects on relative efficiency, and also the assessment of the difference in the MCT of NDV between the subgroups of above- and below- median efficiency would be useful.
Given the difference in results between ROS and ROA it would be possible to consider measuring financial efficiency by some alternative parameter, such as the residual income measure Economic Profit. Regarding normalisation factors, the use of Cost of Sales to normalise the sum of the deals was the most defensible choice, however examining other factors such as revenue, or average R&D expenditure (on the grounds that an acquisition is an alternative means of acquiring a stocked pipeline) in order to undertake a sensitivity analysis could be worthwhile. A further refinement could be to adjust the proportion of Cost of Sales figure for the costs that are related to generic manufacture of pharmaceuticals, however this would prove difficult because this information is not publicly disclosed.
The research has merely touched on the issue of causality by considering the stated intentions of the four mega-mergers that seem to set off ‘aftershocks’ according to some hypotheses of merger dynamics. The next steps are to extend the case-by-case analysis of major acquisitions to establish causality with smaller deals and to analyse further the distribution relating size of deals to their frequency.
157 10. References
Abernethy, M.A., Horne, M., Lillis, A.M., Malina, M.A. and Selto, F.H. (2005) A multi-method approach to building casual performance maps from expert knowledge. Manage. Acc. Res., 16: 135–155.
Accenture (2009) Performance Measurement Frameworks In Board Reports:
Insights Into Leading Practice From The Private And International Public
Sectors, Accenture.
Agrawal, A., Jaffe, J. and Mandelker, G. (1992) The post-merger performance of acquiring firms – a re-examination of an anomaly. Journal of Finance, 47:
1605–1621.
Ahuja, G. and Katila, R. (2001) Technological acquisitions and the innovation performance of acquiring firms – a longitudinal study. Strategic Management
Journal, 22 (3): 197–227.
Ambrosini, V., Bowman, C. and Schoenberg, R. (2010) Should acquiring firms pursue more than one value creation strategy? An empirical test of acquisition performance. British Journal of Management, doi: 10.1111/j.1467-
8551.2010.00693.x
Amit, R. and Livnat, J. (1988) Diversification strategies, business cycles and economic performance. Strategic Management Journal, 9 (2): 99–110.
Amit, R. and Schoemaker, P.J.H. (1993) Strategic assets and organizational rent. Strategic Management Journal, Jan 1993, 14 (1).
Anand, J. and Singh, H. (1997) Asset redeployment, acquisitions and
158 corporate strategy in declining industries. Strategic Management Journal,
Summer Special Issue, 18: 99–118.
Anderson, T.W. and Darling, D.A. (1952) Asymptotic theory of certain
“goodness-of-fit” criteria based on stochastic processes. Annals of
Mathematical Statistics, 23: 193–212.
Angwin, D. (2007) Motive archetypes in mergers and acquisitions (M&A): The implications of a configurational approach to performance. Advances in
Mergers and Acquisitions, 6: 77–105.
Angwin, D. and Vaara, E. (2005) Connectivity in merging organisations: beyond traditional cultural perspectives. Organisation Studies, 26 (10): 1445–
1453.
Association of the British Pharmaceutical Industry (2010) Key Facts. Available from http://www.abpi.org.uk/statistics/section.asp?sect=3
Avrikan, N.K. (1999) The evidence on efficiency gains: The role of mergers and the benefits to the public. Journal of Banking & Finance, 23 (7), July:
991–1013.
Banker, R.D., Chang, H., Janakiraman, S.N. and Konstans, C. (2004) A balanced scorecard analysis of performance metrics. European Journal of
Operational Research, 154, 423–436.
Banker, R.D., Charnes, A. and Cooper, W.W. (1984) Some models for estimating technical and scale inefficiencies in data envelopment analysis.
Management Science, 30 (9), Sept: 1078–1092.
Banker, R.D., Potter, G. and Srinicasan, D. (2000) An empirical investigation of an incentive plan that includes non-financial performance measures. The
Accounting Review, 75 (1): 65–92.
Barber, B.M. and Lyon, J.D. (1997) Detecting long-run abnormal stock returns
– the empirical power and specification of test statistics. Journal of Financial
159 Economics, 43: 341–372.
Barney, J.B. (1991) Firm resources and sustained competitive advantage,
Journal of Management, 17: 99–120.
Barney, J.B. (2001) Is the resource-based "view" a useful perspective for strategic management research? Yes. Strategic Management Journal, 26 (1):
41–56.
Barney, J.B, Ketchen D.J. and Wright M. (2011) The future of resource-based theory: revitalization or decline?, Journal of Management, September, doi:
10.1177/0149206310391805
Beckman, C.M. and Haunschild, P.R. (2002) Network learning – the effects of partners’ heterogeneity of experience on corporate acquisitions.
Administrative Science Quarterly, 47(1): 92–124.
Berger, P.G. and Ofek, E. (1995) Diversification’s effect on firm value. Journal of Financial Economics, 37: 39–65.
Bergh, D.D. (2001) Executive retention and acquisition outcomes – a test of opposing views on the influence of organizational tenure. Journal of
Management, 27: 603–622.
Blackwell, D. (1951) Comparison of experiments. In: Neyman, J. (Ed.)
Proceedings of the Second Berkeley Symposium on Mathematical Statistics and Probability, Berkeley, University of California Press, pp. 93–102.
Boekestein, B. (2009) Acquisitions reveal the hidden intellectual capital of pharmaceutical companies. Journal of Intellectual Capital, 10 (3), 389–400.
Bontis, N., Dragonetty, N.C., Jacobsen, K. and Roos, G. (1999) The knowledge toolbox: a review of the tools available to measure and manage intangible resources. European Management Journal, 17 (4): 391–402.
Bourguignon, A., Malleret, V. and Nørreklit, H. (2004) The American balanced scorecard versus the French tableau de bord: the ideological dimension.
160 Management Accounting Research, 15: 107–134.
Bowerman, B.L., O’Connell, R.T. and Murphree, E.S. (2011) Business
Statistics in Practice, 6th edition, McGraw-Hill.
Braam, G.J.M. and Nijssen, E.J. (2004) Performance effects of using the balanced scorecard: A note on the Dutch experience. Long Range Planning,
37(4), 335–349.
Bradley, M., Desai, A. and Kim, E.H. (1988) Synergistic gains from corporate acquisitions and their division between the stockholders of target and acquiring firms. Journal of Financial Economics, 21: 3–40.
Bresman, H., Birkinshaw, J. and Nobel, R. (1999) Knowledge transfer in international acquisitions, Journal of International Business Studies, 30: 439–
462.
Brooking, A. (1996) Intellectual Capital: Core Assets for the Third Millennium
Enterprise, London, Thompson Business Press.
Brouthers, K.D., van Hastenburg, P. and van den Ven, J. (1998) Long Range
Planning. International Journal of Strategic Management, 31 (3): 347–354.
Brush, T.H. (1996) Predicted change in operational synergy and post- acquisition performance of acquired businesses. Strategic Management
Journal, 17(1): 1–24.
Bruton, G.D., Oviatt, B.M. and White, M.A. (1994) Performance of acquisitions of distressed firms. Academy of Management Journal, 37: 972–989.
Buono, A.F., Bowditch, J.L. and Lewis, J.W. (1985) When cultures collide - the anatomy of a merger. Human Relations, 38: 477–500.
Cañibano, L., García-Ayuso, M. and Sánchez, P. (2000) Accounting for intangibles: a literature review. Journal of Accounting Literature, 19: 102–130.
Cannella, A.A. Jr. and Hambrick, D.C. (1993) Effects of executive departures on the performance of acquired firms. Strategic Management Journal, 14
161 (Special Issue): 137–153.
Capon, N., Hulbert, J. M., Farley, J. U. and Martin, I. E. (1988) Corporate diversity and economic performance - the impact of market specialization.
Strategic Management Journal, 9 (1): 61–74.
Capron, L. (1999) The long-term performance of horizontal acquisitions.
Strategic Management Journal, 20: 987–1018.
Capron, L. and Pistre, N. (2002) When do acquirers earn abnormal returns?
Strategic Management Journal, 23: 781–794.
Carow, K., Heron, R. and Saxton, T. (2004) Do early birds get the returns - an empirical investigation or early-mover advantages in acquisitions. Strategic
Management Journal, 25: 563–585.
Chang, S.J. (1996) An evolutionary perspective on diversification and corporate restructuring - entry, exit, and economic performance during 1981-
89. Strategic Management Journal, 17 (8): 587–611.
Charnes, A., Cooper, W.W. and Rhodes, E. (1978) Measuring the efficiency of decision making units. European Journal of Operational Research, 2: 429–
444.
Chatterjee, S. (1986) Types of synergy and economic value - the impact of acquisitions on merging and rival firms. Strategic Management Journal, 7 (2):
119–139.
Chatterjee, S. (1991) Gains in vertical acquisitions and market power - theory and evidence. Academy of Management Journal, 34 (2): 436–448.
Chatterjee, S. (1992) Sources of value in takeovers: Synergy or restructuring
– Implications for target and bidder firms. Strategic Management Journal, 13:
267–287.
Chatterjee, S., Lubatkin, M.H., Schweiger, D.M. and Weber, Y. (1992) Cultural differences and shareholder value in related mergers - linking equity and
162 human capital. Strategic Management Journal, 13 (5): 319–334.
Chen, T., Chien, C.F., Wang, J.T. and Lin, M.H. (2004) Using DEA to evaluate
R&D performance of the computers and peripherals firms in Taiwan.
International Journal of Business, 9 (4): 347–360.
Clark, K. and Ofek, E. 1994. Mergers as a means of restructuring distressed firms: An empirical investigation. J. Finance. Quant. Anal., 29: 541–565.
Clauset, A., Shalizi C. and Newman M. (2009), Power-law distributions in empirical data, http://arxiv.org/abs/0706.1062v1
Cockburn, I. (2006) Is the Pharmaceutical Industry in a Productivity Crisis?
NBER Innovation Policy and the Economy Conference, Washington DC.
Covin, T.J., Kolenko, T.A., Sightler, K.W. and Tudor, R.K. (1997) Leadership style and post-merger satisfaction. Journal of Management Development, 16
(1): 22–33.
Crook, T.R., Ketchen, D.J. Jr, Combs, J.G. and Todd, S.Y. (2008) Strategic resources and performance: a meta-analysis. Strategic Management Journal,
29 (11): 1141–1154.
Cubbin, J. and Tzanidakis, G. (1998) Techniques for analysing company performance. Business Strategy Review, 9 (4): 37–46.
Datta, D.K. (1991) Organizational fit and acquisition performance: Effects of post-acquisition integration. Strategic Management Journal, 12: 281–298.
Datta, D.K. and Grant, J.H. (1990) Relationships between type of acquisition, the autonomy given to the acquired firm, and acquisition success - an empirical analysis. Journal of Management, 16 (1): 29–44.
Datta, D.K., Pinches, G.E. and Narayanan, V.K. (1992) Factors influencing wealth creation from mergers and acquisitions: A meta-analysis. Strategic
Management Journal, 13: 67–84.
Davis, S. and Albright, T. (2004) An investigation of the effect of Balanced
163 Scorecard implementation on financial performance. Manage. Acc. Res., 15:
135–153.
DeCarolis, D. (2003) Competencies and imitability in the pharmaceutical industry: an analysis of their relationship with firm performance. Journal of
Management, 29 (1): 27–50.
DeCarolis, D. and Deeds, D. (1999) the impact of stocks and flows of organizational knowledge on firm performance: an empirical evaluation of the biotechnology industry. Strategic Management Journal, 20: 953–968.
DeLong, G. and DeYoung, R. (2004) Learning by Observing - Information
Spillovers in the Execution and Valuation of Commercial Bank M&As. Working paper.
Dierickx, I. and Cool, K. (1989) Asset stock accumulation and sustainability of competitive advantage. Management Science, 35 (12): 1504–1511.
DiMasi, J.A. and Grabowski H.G. (2007) Managerial and decision economics.
Manage. Decis. Econ., 28: 469–479.
DiMasi, J.A., Hansen, R.W. and Grabowski, H.G. (2003) The price of innovation: new estimates of drug development costs. Journal of Health
Economics, 22: 151–185.
Dyson, R.G. and Shale, E.A. (2010) Data envelopment analysis, operational research and uncertainty. Journal of the Operational Research Society, 61:
25–34.
Eckbo, B.E. (1983) Horizontal mergers, collusion, and stockholder wealth,
Journal of Financial Economics (JFE), 11: 241–273.
Edvinsson, L. and Malone, M.S. (1997) Intellectual Capital: The Proven Way to Establish Your Company’s Real Value By Measuring Its Hidden Values.
London, Piatkus.
Edvinsson, L. and Sullivan, P. (1996) Developing a model of managing
164 intellectual capital. European Management Journal, 4 (4): 356–364.
Eilat, H., Golany, B. and Shtub, A. (2006) Constructing and evaluating balanced portfolios of R&D projects with interactions: A DEA based methodology. European Journal of Operational Research, 172: 1018–1039.
Eilat, H., Golany, B. and Shtub, A. (2008) R&D project evaluation: An integrated DEA and balanced scorecard approach. Omega International
Journal of Management Science, 36: 895–912.
Elgers, P.T. and Clark, J.J. (1980) Merger types and stockholder returns:
Additional evidence. Financial Management, 9: 66–72.
Enhanced Business Reporting Consortium (2006) The Enhanced Business
Reporting Framework for Private Companies, July, Washington, DC 20036.
Eun, C.S., Kolodny, R. and Scheraga, C. (1996) Cross-border acquisitions and shareholder wealth: Tests of synergy and internalization hypotheses.
Journal of Banking and Finance, 20: 1559–1582.
Fama, E.F. (1970) Efficient capital markets: a review of theory and empirical work. The Journal of Finance, (25) 2. Papers and Proceedings of the Twenty-
Eighth Annual Meeting of the American Finance Association New York, N.Y.
December, 28–30, 1969 (May, 1970), pp. 383–417.
Fama, E.F., Fisher, L., Jensen, M. and Roll, R. (1969) The adjustment of stock prices to new information. International Economic Review, 10: 1–21.
Farmer J.D. and Geanakoplos J., (2008), Power laws in economics and elsewhere. Unpublished book chapter, http://www.elautomataeconomico.com.ar/download/papers/Farmer- powerlaw3.pdf
Farrell, M.J. (1957) The measurement of productive efficiency. Journal of the
Royal Statistical Society. Series A (General), 120 (3): 253–290.
Feea, C E. and Thomas, S. (2004) Sources of gains in horizontal mergers -
165 evidence from customers, supplier, and rival firms. Journal of Financial
Economics, 74 (3): 423–460.
Ferrier, G. and Valmanis, V. (2004) Do mergers improve hospital productivity?
Journal of the Operational Research Society, 55: 1071–1080.
Fisher R.A. (1925) Statistical Methods for Research Workers, Edinburgh,
Oliver & Boyd.
Fisher R.A. (1956) Statistical methods and scientific induction. Journal of the
Royal Statistical Society, Series B, 17, 69–78.
Foss, N.J. (1997) The Resource-Based Perspective: An Assessment and
Diagnosis of Problems: 1-33. Copenhagen Business School, Department of
Industrial Economics and Strategy: DRUID Working Papers series 97-1.
Foss, N.J. and Knudsen, T. (2003) The resource-based tangle: towards a sustainable explanation of competitive advantage. Managerial and Decision
Economics, 24: 291–307.
Fowler, K.L. and Schmidt D.R. (1989) Determinants of tender offer post- acquisition financial performance. Strategic Management Journal, 10 (4):
339–350.
Franks, J., Harris, R. and Titman, S. (1991) The postmerger share-price performance of acquiring firms. Journal of Financial Economics, 29, 81–96.
Garcia-Valderrama, T., Mulero-Mendigorri E. and Revuelta-Bordoy, D. (2009)
Relating the perspectives of the balanced scorecard for R&D by means of
DEA. European Journal of Operational Research, 196 (3), 1 August: 1177–
1189.
Grabowski H., Vernon J. and DiMasi J., (2002), Returns on research and development for 1990s new drug introductions, Pharmacoeconomics, 20
(Suppl. 3): 11–29.
166 Graves, S.B. and Langowitz, N.S. (1993) Innovative productivity and returns to scale in the pharmaceutical industry. Strategic Management Journal, 14 (8):
593–605.
Grosskopf S. (1996), Statistical inference and nonparametric efficiency: A selective survey. Journal of Productivity Analysis, 7, (2–3): 161–176.
Haleblian, J. and Finkelstein, S. (1999) The influence of organizational acquisition experience on acquisition performance: A behavioral perspective.
Administrative Science Quarterly, 44: 29–56.
Hall, R. (1992) The Strategic analysis of intangible resources. Strategic
Management Journal, 13 (2), 135–144.
Hambrick, D.C. and Cannella, A.A. Jr. (1993) Relative standing: a framework for understanding departures of acquired executives. Academy of
Management Journal, 36, 733-762.
Hamilton, G. (2005) The Pharmaceutical Market Outlook to 2015, Business
Insights.
Harris, R.S. and Ravenscraft, D. (1991) The role of acquisitions in foreign direct investment - evidence from the U.S. stock market. Journal of Finance,
46 (3): 825–844.
Harrison, J.S., Hitt, M.A., Hoskisson, R.E. and Ireland, R D. (1991) Synergies and post-acquisition performance - differences versus similarities in resource allocations. Journal of Management, 17 (1): 173–190.
Harrison, J.S., Oler, D. and Allen, M.R. (2005) Event Studies and the
Importance of Longer Term Measures in Assessing the Performance
Outcomes, Working paper.
Hashimoto, A. and Haneda, S. (2008) Measuring the change in R&D efficiency of the Japanese pharmaceutical industry. Res. Policy, 37 (10):
1829–1836.
167 Haspeslagh, P.C. and Jemison, D.B. (1991) Managing Acquisitions: Creating
Value Through Corporate Renewal, New York: The Free Press.
Hayward, M.L.A. (2002) When do firms learn from their acquisition experience? Evidence from 1990-1995. Strategic Management Journal, 23
(1): 21–39.
Hayward, M.L.A. and Hambrick, D.C. (1997) Explaining the premium paid for large acquisitions: Evidence of CEO hubris. Administrative Science Quarterly,
42: 103–127.
Healy P., Palepu, K. and Ruback, R. (1997) Which takeovers are profitable: strategic or financial? Sloan Management Review, 38:45–57.
Heron, R. and Lie, E. (2002) Operating performance and the method of payments in takeovers. Journal of Financial and Quantitative Analysis, 37 (1):
137–156.
Higgins M.J. and Rodriguez D. (2006) The outsourcing of R&D through acquisitions in the pharmaceutical industry, Journal of Financial Economics,
80 (2), May: 351–383.
Hitt, M., Harrison, J., Ireland, R.D. and Best, A. (1998) Attributes of successful and unsuccessful acquisitions of US firms. British Journal of Management, 9
(2): 91–114.
Hitt, M.A., Hoskisson, R.E., Ireland, R.D. and Harrison, J.S. (1991) Effects of acquisitions on R&D inputs and outputs. Academy of Management Journal, 34
(3): 693–706.
Hitt, M.A., Hoskisson, R.E., Johnson, R. and Moesel, D. (1996) The market for corporate control and firm innovation. Academy of Management Journal, 39:
1084–1109.
H.M. Treasury, Cabinet Office, National Audit Office, Audit Commission &
168 Office for National Statistics (2001) Choosing the Right FABRIC: A Framework for Performance Information.
Holl, P. and Kyriazis, D. (1997) Wealth creation and bid resistance in UK takeover bids. Strategic Management Journal, 18: 483–498.
Holmström, B. (1979) Moral hazard and observability. The Bell Journal of
Economics, 10 (1), Spring: 74–91.
Homburg, C. and Bucerius, M. (2006) Is speed of integration really a success factor of mergers and acquisitions? An analysis of the role of internal and external relatedness. Strategic Management Journal, 27: 347–67.
Hoskisson, R.E., Hitt, M.A., Johson, R.A. and Moesel, D.D. (1993) Construct validity of an objective (entropy) categorical measure of diversification strategy. Strategic Management Journal, 14 (3): 215–235.
Hunt, J.W. (1990) Changing pattern of acquisition behaviour in takeovers and the consequences for acquisition processes. Strategic Management Journal,
11 (1): 69–77.
Ittner, C. and Larcker, D.F. (1998) Are nonfinancial measures leading indicators of financial performance? An analysis of customer satisfaction.
Journal of Accounting Research, 36 (Supplement): 1–36.
Ittner, C.D. and Larcker, D.F. (2003) Coming up short on nonfinancial performance measurement. Harvard Business Review (November): 88–95.
Ittner, C.D., Larcker, D.F. and Randall, T. (2003) Performance implications of strategic performance measurement in financial services firms. Accounting,
Organizations and Society, 28: 715–741.
Jensen, E.J. (1987) Research expenditures and the discovery of new drugs.
Journal of Industrial Economics, 36: 83–95.
Jensen, M.C. and Ruback, R.S. (1983) The market for corporate control: The
169 scientific evidence. Journal of Financial Economics, 11: 5–50.
Kaplan, R.S. (2010) Interview. See http://www.bnet.com/blog/harvard/risk- management-the-balanced-scorecards-missing-piece/6050, on 12 March
2010.
Kaplan, R.S. and Norton, D.P. (1992) The balanced scorecard - Measures that drive performance. Harvard Business Review, (January-February): 71–
79.
Kaplan, R.S. and Norton, D.P. (1996) Using the balanced scorecard as a strategic management system. Harvard Business Review, (January/February):
75–85.
Kapoor R. and Lim K. (2007) The impact of acquisitions on the innovation performance of inventors at semiconductor companies. Academy of
Management Journal, 50 (5): 1133–1155.
King, D.R., Dalton, D.R., Daily, C.M. and Covin, J.G. (2004) Meta-analyses of
PAP: indications of unidentified moderators. Strategic Management Journal,
25 (2), Feb: 187–200.
Kitching, J. (1967) Why do mergers miscarry? Harvard Business Review, 45
(6): 84–102.
Kraaijenbrink J., Spender, J-C. and Groen A.J. 2010 The resource-based view: A review and assessment of its critiques, Journal of Management, 36
(1): 349–372.
Kramer, J.K. and Pushner, G. (1997) An empirical analysis of economic value added as a proxy for market value added. Financial Practice and Education,
Spring/ Summer: 41–49.
Krishnan, H.A., Miller, A. and Judge, W.Q. (1997) Diversification and top management team complementarity. Strategic Management Journal, 18 (5):
361–374.
170 Kroll, M., Wright, P., Toombs, L. and Leavell, H. (1997) Form of control: A critical determinant of acquisition performance and CEO rewards. Strategic
Management Journal, 18: 85–96.
Krug, J.A. and Hegarty, W.H. (2001) Predicting who stays and leaves after an acquisition – a study of top managers in multinationals. Strategic Management
Journal, 22 (2): 185–196.
Kumbhakar S.C. (1988) Estimation of input-specific technical and allocative inefficiency in stochastic frontier models. Oxford Economic Papers, New
Series, 40 (3), Sept: 535–549.
Kusewitt, J.B. (1985) An exploratory study of strategic acquisitions factors relating to performance. Strategic Management Journal, 6: 151–169.
Lackman C., Saban K. and Lanasa J. (2000), The contribution of market intelligence to tactical and strategic business decisions, Marketing
Intelligence & Planning,18/1
Lahey, K.E. and Conn, R.L. (1990) Sensitivity of acquiring firms' returns to alternative model specification and disaggregation. Journal of Business and
Accounting, 17: 421–439.
Larsson, R. and Finkelstein, S. (1999) Integrating strategic, organizational, and human resource perspectives on mergers and acquisitions - a case survey of synergy relatization. Organization Science, 10 (1): 1–26.
Lawrie, G. and Cobbold, I. (2004) Third-generation balanced scorecard: evolution of an effective strategic control tool. International Journal of
Productivity and Performance Management, 53 (7): 611–623.
Lebas, M. and Euske, K. (2002) A conceptual and operational delineation of performance. In Neely, A. (Eds) Business Performance Measurement: Theory and Practice, Cambridge University Press, Cambridge
Lee, H. and Park, Y. (2005) An international comparison of R&D efficiency:
171 DEA approach. Asian Journal of Technology Innovation, 13 (2): 207–221.
Lehman, E.L. (1993) The Fisher, Neyman-Pearson theories of testing hypotheses: One theory or two? Journal of the American Statistical
Association, 88: 1242–1249.
Leitner, K. and Warden, C. (2004) Managing and reporting knowledge-based resources and processes in research organisations: specifics, lessons learned and perspectives. Management Accounting Research, 15: 33–51.
Leitner, K.H., Linzatti, M.S., Stowasser, R. and Wagner, K. (2005) Data envelopment analysis as method for evaluating intellectual capital. Journal of
Intellectual Capital, 6 (4): 528–543.
Lev, B. (2001) Intangibles: Management, Measurement, and Reporting,
Washington, DC: The Brookings Institution.
Linton, J.D., Walsh, S.T. and Morabito, J. (2002) Analysis, ranking, selection of R&D projects in a portfolio. R&D Management, 32: 139–148.
Lorie, J.H. and Halpern, P.J. (1970) Conglomerates: The rhetoric and the evidence. Journal of Law and Economics, 13: 149–166.
Loughran, T. and Vijh, A. M. (1997) Do long-term shareholders benefit from corporate acquisitions. Journal of Finance, 52 (5): 1765–1790.
Lubatkin, M. (1987) Merger strategies and stockholder value. Strategic
Management Journal, 8: 39–53.
Lubatkin, M., Srinivasan, N. and Merchant, H. (1997) Merger strategies and shareholder value during times of relaxed antitrust enforcement - the case of large mergers during the 1980s. Journal of Management, 23: 59–81.
Malina, M.A. and Selto, F.H. (2004) Choice and change of measures in performance measurement models. Manage. Acc. Res., 15: 441–469.
Malo, J.L. (1995) Les tableaux de bord comme signes d’une gestion et d’une compatabilité à la Française. In: Mélanges en L’honneur du Professeur
172 Claude Perochon, Paris: Foucher, pp. 357–376.
Mann, H.B. and Whitney, D.R. (1947) On a test of whether one of two random variables is stochastically larger than the other. Ann. Math. Statist., 18: 5060.
Markides, C.C. and Ittner, C.D. (1994) Shareholder benefits from corporate international diversification - evidence from U.S. international acquisitions.
Journal of International Business Studies, 25(2): 343–366.
Marks, M.L. and Mirvis, P.H. (1998) Joining Forces: Making One Plus One
Equal Three in Mergers, Acquisitions, And Alliances, San Francisco: Jossey-
Bass Publishers.
Marr, B. and Schiuma, G. (2001) Measuring and managing intellectual capital and knowledge assets in new economy organisations. In Bourne, M. (Ed.),
Handbook of Performance Measurement, London: Gee.
Marr, B., Gray, D. and Schiuma, G. (2004) Measuring intellectual capital – what, why, and how? In Bourne, M. (Ed.), Handbook of Performance
Measurement, London: Gee.
Martin, J.D. and Sayrak, A. (2003) Corporate diversification and shareholder value: a survey of recent literature. Journal of Corporate Finance, 9: 37–57.
McCullagh P. and Nelder J.A. (1989) Generalized Linear Models, London:
Chapman and Hall.
Meyer, M.S. (2001) Patent citation analysis in a novel field of technology: An exploration of nano-science and nano-technology. Computer Science, 51 (1):
163–.183.
Mitzenmacher, M. (2001) A brief history of generative models for power law and lognormal distributions. Internet Mathematics, 1 (2): 226–251.
Moeller, S.B., Schlingemann, F.P. and Stulz, R. (2004) Firm size and the gains from acquisitions. Journal of Financial Economics, 73 (2): 201–228.
Molloy J.C., Chadwick C., Ployhart R.E. and Golden S.J. (2011) Making
173 intangibles "tangible" in tests of resource-based theory: a multidisciplinary construct validation approach. Journal of Management, September, doi:10.1177/0149206310394185
Montgomery, C.A. and Wilson, V.A. (1986) Mergers that last - a predictable pattern. Strategic Management Journal, 7 (1): 91–96.
Morck, R., Shleifer, A. and Vishny, R. (1988) Management ownership and market valuation: an empirical analysis. Journal of Financial Economics, 20
(1/2): 293–315.
Morosini, P., Shane, S. and Singh, H. (1998) National cultural distance and cross-border acquisition performance. Journal of International Business
Studies, 29 (1): 137–158.
Murray, E.J. and Richardson, P.R. (2000) Shared understanding of the critical few: a parameter of strategic planning effectiveness. International Journal of
Business Performance Management, 2 (1/2/3): 5–16.
Neely, A. (2000) Performance measurement: theory and practice.
International Journal of Business Performance Management, 2 (1/2/3): 1–4.
Neely, A., Adams, C. and Kennerley, M. (2002) The Performance Prism - The
Scorecard for Measuring and Managing Business Success, Financial Times
Prentice Hall.
Neely, A., Gregory, M. and Platts, K. (1995) Performance measurement systems. International Journal of Operations and Production Management, 15
(4): 80–116.
Neyman, J. and Pearson, E. (1933) On the problem of the most efficient tests of statistical hypotheses. Philosophical Transactions of the Royal Society of
London, Series A, 231: 289–337.
OFWAT (2005) A Consultation Paper on Monitoring Companies’
Performance, Birmingham, AMP 4, June return 2005–10.
174 Oral, M., Kettani, O. and Lang, P. (1991) A methodology for collective evaluation and selection of industrial R&D projects. Management Science, 37:
7.
ORR (2004) Consultation for a Balanced Scorecard for Network Rail,
Network Rail Monitor, November, 137.
Ozgen, H. and Ozcan, Y. A. (2000) Do mergers enhance the performance of hospital efficiency? Journal of the Operational Research Society, 51: 801–
811.
Palich, L.E., Cardinal, L.B. and Miller, C.C. (2000) Curvilinearity in the diversification performance linkage - an examination of over three decades of research. Strategic Management Journal, 21 (2): 155–174.
Pangarkar, N. (2004) The impact of market cycle on the performance of
Singapore acquirers. Strategic Management Journal, 25 (12): 1209–1216.
Papadakis, V.N. and Thanos, T. (2010) Ioannis, measuring the performance of acquisitions: an empirical investigation using multiple criteria. British
Journal of Management, 21 (4): 859–873.
Papalexandris, A., Ioannou, G., Prastacos, G. and Soderquist, K.E. (2005) An integrated methodology for putting the balanced scorecard into action. Eur
Mngt J, 23 (2): 214–227.
Park, J., Morel, B. and Madhaven, R. 2010 Big Surf: Self-Organised Criticality in M&A Waves, SSRN working paper.
Pennings, J.M., Barkema, H. and Douma, S. (1994) Organizational learning and diversification. Academy of Management Journal, 37 (3): 608–640.
Penrose, E. (1959) The Theory of the Growth of the Firm, Oxford: Oxford
University Press.
Peteraf, M.A. (1993) The cornerstones of competitive advantage: a resource-
175 based view. Strategic Management Journal, 14: 179–191.
Peteraf, M.A. and Barney, J.B. (2003) Unravelling the resource-based tangle.
Managerial and Decision Economics, 24 (4): 309–323.
Pharmaceutical Research and Manufacturers of America (2010)
Pharmaceutical Industry Profile 2010, Washington, DC: PhRMA.
Pharmalive (2007) Top 50 Pharmaceutical Company Pipelines, Los Angeles:
Canon Communications LLC.
Pharmaventures (2003) Valuing and Structuring Pharmaceutical Licensing
Agreements, page 8, Pharmaventures, Oxford, UK.
Pozen, David E. (2005) The mosaic theory, national security, and the
Freedom of Information Act. Yale Law Journal, 115: 628–679.
Prahalad, C.K. and Hamel, G. (1990) the core competence of the organization. Harvard Business Review, May-June: 79–90.
Priem, R.L. and Butler J.E. (2001a) Is the resource-based "view" a useful perspective for strategic management research. Academy of Management
Review, 26 (1): 22–40.
Priem, R.L. and Butler J.E. (2001b) Tautology in the resource-based view and the implications of externally determined value: further comments. Academy of Management Review, 26 (1): 57–66.
Puranam, P., Singh, H. and Zollo, M. (2006) Organizing for innovation:
Managing the coordination-autonomy dilemma in technology acquisitions.
Academy of Management Journal, 49 (2): 263 - 280.
Ramaswamy, K. (1997) The performance impact of strategic similarity in horizontal mergers - evidence from the U.S. banking industry. Academy of
Management Journal, 40 (3): 697–715.
Ravenscraft, D.J. and Scherer, F.M. (1987) Life after takeover. Journal of
Industrial Economics, 36 (2): 147–156.
176 Reid, S.R. (1968) Mergers, Managers and the Economy, New York: McGraw-
Hill.
Rockart, J.F. (1979) Chief executives define their own data needs. Harvard
Business Review, (2): 81–93.
Roos, J., Roos, G., Dragonetti, N.C. and Edvinsson, L. (1997) Intellectual
Capital: Navigating the New Business Landscape, London: Macmillan.
Rucci, A.J., Kirn, S.P. and Quinn, R.T. (1998) The employee-customer-profit chain at Sears. Harvard Business Review, 76 (1): 82–97.
Rumelt, R.P. (1984) Toward a strategic theory of the firm. In R. Lamb (ed.),
Competitive Strategic Management, Englewood Cliffs, NJ: Prentice Hall, pp.
556– 570.
Schoenberg, R. (2006) Measuring the performance of corporate acquisitions: an empirical comparison of alternative metrics. British Journal of
Management, 17 (4): 361–370.
Schoenberg, R. and Reeves, R. (1999) What determines acquisition activity within an industry? European Management Journal, 17 (1): 93–98(6).
Schumpeter, J. A. (1950) Capitalism, Socialism, and Democracy, 3rd edn,
New York: Harper & Row.
Schweiger, D.M. and Denisi, A.S. (1991) Communication with employees following a merger - a longitudinal field experiment. Academy of Management
Journal, 34 (1): 110–135.
Seth, A. (1990a) Value creation in acquisitions: A re-examination of performance. Strategic Management Journal, 11: 99–116.
Seth, A. 1990b. Sources of value creation in acquisitions: An empirical investigation. Strategic Management Journal, 11: 431–447.
177 Seth, A., Song, K.P. and Pettit, R. (2000) Synergy, managerialism or hubris?
An empirical examination of motives for foreign acquisitions of U.S. firms.
Journal of International Business Studies, 31 (3): 387–405.
Seth, A., Song, K.P. and Pettit, A.R. (2002) Value creation and destruction in cross-border acquisitions. Strategic Management Journal, 23: 921–940.
Shahrur, H. (2005) Industry structure and horizontal takeovers - analysis of wealth effects on rivals, suppliers, and corporate customers. Journal of
Financial Economics, 76 (1): 61–98.
Shanley, M.T. and Correa, M.E. (1992) Agreement between top management teams and expectations for post-acquisition performance. Strategic
Management Journal, 13(4): 245–266.
Sharma, S. and Thomas, V.J. (2008) Inter-country R&D efficiency analysis: An application of data envelopment analysis. Scientometrics, 76 (3): 483–501.
Shelton, L.M. (1988) Strategic business fits and corporate acquisition - empirical evidence. Strategic Management Journal, 9 (3): 279–287.
Sheskin D. (2011), Handbook of Parametric and Nonparametric Statistical
Procedures, Fifth edition, Chapman & Hall/CRC.
Singh, H. and Montgomery, C.A. (1987) Corporate acquisition strategies and economic performance. Strategic Management Journal, 8: 377–387.
Sirower, M.L. (1997) The Synergy Trap: How Companies Lose the Acquisition
Game, New York: The Free Press.
Slusky, A.R. and Caves, R.E. (1991) Synergy, agency, and the determinants of premia paid in mergers. Journal of Industrial Economics, 39 (3): 277–296.
Smith, K.V. and Schreiner, J.C. (1969) A portfolio analysis of conglomerate. diversification. Journal of Finance, 13: 413–428.
Speckbacher, G., Bischofb, J. and Pfeiffer, T. (2003) A descriptive analysis on the implementation of Balanced Scorecards in German-speaking countries.
178 Management Accounting Research, 14 (4): 361–388.
Stankevicien J. and Svidersk T. (2010) Developing a performance measurement system integrating economic value added and the balanced scorecard in pharmaceutical company, 6th International Scientific Conference,
May 13–14, Vilnius, Lithuania, Business and Management, doi:10.3846/bm.2010.033
Stewart, T.A. (1997) Intellectual Capital: The New Wealth of Organizations,
New York, NY: Doubleday/Currency.
SubbaNarasimha, P.N., Ahmad, S. and Mallya, S.N. (2003) Technological knowledge and firm performance of pharmaceutical firms. Journal of
Intellectual Capital, 4 (1): 20–33.
Sveiby, K.E. (1997) The New Organizational Wealth: Managing and
Measuring Knowledge-based Assets, San Francisco, CA: Barrett-Kohler.
Sweeny, K. (2002) Technology Trends in Drug Discovery and Development:
Implications for the Development of the Pharmaceutical Industry in Australia,
Working paper No. 3.
Sykes, D. (1999) Pharmaceutical Mergers and Acquisitions, Business
Insights.
Tavares, G. (2002) A Bibliography of Data Envelopment Analysis (1978-
2001), Rutcor Research Report Piscataway, New Jersey: Rutgers University.
Teece, D., Pisano, G. and Shuen, A. 1997. Dynamic capabilities and strategic management. Strategic Management Journal, 18 (7): 509–533.
Thakor, A.V. (1999) Information technologies and financial services consolidation. Journal of Banking and Finance, 23: 697–700.
Travlos, N.G. (1987) Corporate takeover bids, methods of payment, and bidding firms' stock returns. Journal of Finance, 42: 943–963.
179 Travlos, N.G. and Waegelein, J.F. (1992) Executive compensation, method of payment and abnormal returns to bidding firms at takeover announcements.
Managerial & Decision Economics 13(6), Nov./Dec: 493–501.
Venkatraman, N. and Grant, J.H. (1986) Construct measurement in organizational strategy research: A critique and proposal. Academy of
Management Review, 11 (1): 71–87.
Vermeulen, F. and Barkema, H. (1996) Learning through acquisition.
Academy of Management Journal, 44 (3): 457–476.
Walker, M. (2000) Corporate takeovers, strategic objectives, and acquiring- firm shareholder wealth. Financial Management, 29: 53–66.
Walsh, J.P. (1988) Top management turnover following mergers and acquisitions. Strategic Management Journal, 9: 173–184.
Walsh, J.P. (1989) Doing a deal - merger and acquisition negotiations and their impact upon target company top management turnover. Strategic
Management Journal, 10 (4): 307–322.
Wansley, J., Lane, W. and Yang, H. (1983) Abnormal returns to acquired firms by type of acquisition and method of payment. Financial Management, 12:
16–22.
Weber, Y. (1996) Corporate cultural fit and performance in mergers and acquisitions. Human Relations, 49 (9): 1181–1203.
Welch, B.L. (1947) The generalization of "Student's" problem when several different population variances are involved. Biometrika, 34 (1–2): 28–35.
Wernerfelt, B. (1984) A resource-based view of the organisation. Strategic
Management Journal, 5: 171–180.
Weston, J.F. and Mansinghka, S.K. (1971) Tests of the efficiency performance in conglomerate firms. Journal of Finance, 26: 919–936.
180 Williamson, O.E. (1999) Strategy research: governance and competence perspectives. Strategic Management Journal, 20: 1087–1108.
Wong, Y.H.B. and Beasley, J.E. (1990) Restricting weight flexibility in data envelopment analysis. J Opl Res Soc, 41: 829–835.
Worthington, A.C. (2004) Determinants of merger and acquisition activity in
Australian cooperative deposit-taking institutions. Journal of Business
Research, 57 (1): 47–57.
Yeoh, P. and Roth, K. (1999) An empirical analysis of sustained advantage in the US pharmaceutical: impact of firm resources and capabilities. Strategic
Management Journal, 20 (7): 637–653.
Zimmerman, D.W. (1998) Invalidation of parametric and nonparametric statistical tests by concurrent violation of two assumptions. The Journal of
Experimental Education, 67: 55–68.
Zollo, M. and Meier, D. (2008) What is M&A performance? Academy of
Management Perspectives, 22: 55–77.
Zollo, M. and Reuer, J.J. (2005) Experience Spillovers Across Corporate
Development Activities, Working paper.
Zollo, M. and Singh, H. (2004) Deliberate learning in corporate acquisitions - post-acquisition strategies and integration capability in U.S. bank mergers.
Strategic Management Journal, 25: 1233–1256.
181 A. R&D Data
A.1 Introduction
The R&D data used in the models are a combination of primary data and calculated data. The origins of the primary data and the adjustments are detailed in this Appendix.
A.2 Historical R&D Data
The historical R&D expenditure in $million unadjusted for inflation is shown in
Table A.1 for the years from 2005 to 2001. These data are drawn from the same source to enhance compatibility. If this source did not have data for a particular firm or for a particular year, then the cells were left blank in Table
A.1.
Table A.1 Historical R&D Data in US$million Actual
Firm R&D 2005 R&D 2004 R&D 2003 R&D 2002 R&D 2001
Abbt 1821 1697 1624 1475 1492
Akzo 681 640 705 747 666
Alco 422 390 350 324 290
Alle 391 346 764 233 228
Amg 2314 2028 1655 1167 865
Astel 1214 1091 1221 562 547
182 Astr 3379 3467 3012 3069 2687
Baus 178 163 150 128 122
Baxt 533 517 553 501 426
Baye 1188 1300 1555 1727 1973
Biog 748 686 534
BMS 2746 2500 2279 2206 2157
Boeh 1693 1534 1464 1623 1269
Ceph 125 134 102 82 64
Chug 424 408 369 411 405
CSL
Daic 1357 1241
Dain 253 149 136 130 112
Eisa 797 669 590 510 470
EliL 3026 2691 2350 2149 2235
Fore 410 294 234 205 158
Gene 1262 948 722 623 526
Genz 503 392 335 308 264
Gile
183 GSK 5709 5286 5215 5279 4826
John 6312 5203 4684 3957 3591
King
Kyow 239 207 2111 232 213
Lund 300 296 322 262 257
Merc 3848 4010 3280 2667 2456
Merk 721 612 630 621 596
Mits 410 431 432 413 293
Nova 4846 4171 3729 2843 2528
Novo 848 726 676 659 662
Nyco
Pfiz 7442 7684 7487 5208 4776
Roch 4579 4137 3825 3417 3125
Sano 5034 4935 1638 1516 1283
Schr 1865 1697 1469 1425 1312
Shio 276 251 255 267 262
Shir
Solv 437 366 354 335 276
184 Tais 197 198 207 252 275
Take 1417 1156 1046 972 792
Teva 369 338 214 165 107
UCB 642 454 264 268 211
Wats
Wyet 2749 2461 2094 2080 1870
A.3 Historical Sales Data
The estimation of a single parameter for historical R&D also requires the revenue statistics for the years 2001 to 2005. These data are given below, unadjusted for inflation. Blank cells indicate unavailable data.
Table A.2 Historical Revenue Data in $million Actual
Firm Rev. 2005 Rev. 2004 Rev. 2003 Rev. 2002 Rev. 2001
Abbt 22338 19680 17280 15280 13919
Akzo 4381 4197 4419 4990 5034
Alco 4369 3914 3407 3009 2748
Alle 2319 2946 1755 1385 1142
Amg 12430 10550 8356 5523 4016
185 Astel 7434 7195 3851 3778 3544
Astr 23950 21426 18849 17841 16222
Baus 2335 2232 2020 1817 1666
Baxt 9849 9509 8904 8099 7342
Baye 11738 10031 11044 11667 13309
Biog 2423 2212 1852
BMS 19207 19380 18653 16208 16612
Boeh 11870 10155 9190 9436 8333
Ceph 1646 1641 1458 1223 1160
Chug 2773 2497 1972 2012 1794
CSL
Daic 6707 6552
Dain 1646 1276 1258 1272 1200
Eisa 4957 4369 4077 3777 3466
EliL 14645 13858 12583 11078 11543
Fore 2962 3160 2680 2246 1602
Gene 6633 4621 3300 2584 2044
Genz 2735 2201 1714 1329 1224
186 Gile
GSK 39430 36383 38356 38614 37928
John 50514 47348 41862 36298 32317
King
Kyow 1769 1831 1813 1702 1691
Lund 1513 1623 1658 1583 1277
Merc 22012 22939 22486 21446 21199
Merk 4848 4943 6849 6994 7259
Mits 2019 2002 2012 2390 1957
Nova 32212 28247 24864 20877 18762
Novo 5631 4842 4363 4148 3901
Nyco
Pfiz 51298 52516 44736 32294 29024
Roch 28502 23695 25058 23640 23407
Sano 33999 31370 29019 9272 8077
Schr 9508 8272 8334 10180 9762
Shio 1653 1677 1681 2397 3504
Shir
187 Solv 2826 2172 2281 2319 2202
Tais 2320 2388 2448 2343 2320
Take 9184 8295 8088 7605
Teva 5250 4790 3276 2519 2077
UCB 2941 2368 1838 1854 1793
Wats
Wyet 18776 17358 15851 14854 13984
A.4 Current R&D & Composite R&D Data
Column 2 of Table A.3 shows the R&D expenditure for 2006 and Column 3 shows the R&D expenditure for 2005, drawn from a source which allows direct comparison between those two years. In most cases the 2005 data are the same as shown in Column 2 of Table A.1 (for example ‘Abbt’ has an expenditure of $1821 million in both cases) however there can be small differences (for example ‘Akzo’ has an expenditure of $687m in the data shown in Table A.3 but an expenditure of $681m in Table A.1).
It is necessary to adjust the 2005 R&D expenditure shown in Column 3 of
Table A.3 for 2005 for two effects:
the rate of inflation of the US dollar from 2005 to 2006;
the historic trend in R&D from 2001 to 2005 as shown in Table A.1.
188 The first adjustment requires the inflation of the 2005 data by 3%4 to reflect
2006 prices. The second adjustment is more complex and is described below.
The data in Tables A.1 and A.2 (which are comparable between years) are used to express R&D expenditure as a percentage of Revenue for each year.
The ratio of the percentage in 2005 to the average ratio for the years from
2001 to 20045 is then used to apply an adjustment factor (reflecting the historic trend in R&D as a percentage of Revenue) to the 2005 R&D expenditure in
Column 3 of Table A.3. This calculation of the historic adjustment factor is shown for each firm in Columns 4, 5 and 6 respectively of Table A.3.
The final column, Column 7 in Table A.3, shows the data used as in input to the DEA model for historic DEA, which is the product of the Historic
Adjustment to reflect historic R&D expenditure as a percentage of Revenue and the changing price levels from 2005 to 2006.
The adjustment can be expressed algebraically as:
Historic R&D DEA Input equals
Comparable 2005 R&D Expenditure
times Inflation adjustment to 2006 price levels
times average R&D as % sales 2001 to 2004 average R&D as % sales in 2005
The results of the calculation are shown in Table A.3.
4 http://data.bls.gov/cgi-bin/cpicalc.pl?cost1=1000&year1=2005&year2=2006
5 The R&D expenditure for the year 2005 is excluded from the calculation of the average, so as to avoid the circularity of adjusting a figure for a historic trend that includes the figure itself.
189 Table A.3 Current & Composite R&D Data in $million Actual
Mean DEA
R&D R&D/Rev R&D/Rev Input
2005 for 2005 ’04 – ’01 R&D (comp. Historic Historic
2006 to ’06) % % Adjustment R&D
Abbt 2255 1821 8.15% 9.60% 117.74% 2208
Akzo 741 687 15.54% 14.85% 95.54% 676
Alco 512 422 9.66% 9.25% 95.81% 416
Alle 1056 388 16.86% 23.02% 136.51% 546
Amg 3366 2314 18.62% 20.42% 109.71% 2615
Astel 1435 1214 16.33% 19.29% 118.15% 1477
Astr 3902 3379 14.11% 16.48% 116.82% 4066
Baus 197 178 7.62% 7.27% 95.42% 175
Baxt 614 533 5.41% 5.91% 109.19% 599
Baye 1791 1048 10.12% 14.17% 139.97% 1511
Biog 718 748 30.87% 29.92% 96.93% 747
BMS 3067 2746 14.30% 12.93% 90.43% 2558
Boeh 1977 1709 14.26% 15.87% 111.24% 1958
190 Ceph 403 355 7.59% 6.85% 90.15% 330
Chug 467 428 15.29% 19.51% 127.62% 563
CSL 161 136 100.00% 140
Daic 1459 1357 20.23% 18.94% 93.62% 1308
Dain 350 253 15.37% 9.71% 63.19% 165
Eisa 926 797 16.08% 12.99% 80.78% 663
EliL 3129 3026 20.66% 19.21% 92.99% 2898
Fore 941 410 13.84% 9.26% 66.87% 282
Gene 1773 1262 19.03% 27.43% 144.16% 1874
Genz 650 503 18.39% 20.52% 111.60% 578
Gile 384 278 100.00% 286
GSK 6373 5781 14.48% 13.63% 94.14% 5605
John 7125 6462 12.50% 11.05% 88.41% 5885
King 254 263 100.00% 271
Kyow 268 264 13.51% 38.49% 284.91% 775
Lund 329 300 19.83% 18.58% 93.72% 290
Merc 4783 3848 17.48% 14.02% 80.21% 3179
Merk 772 728 14.87% 9.67% 65.00% 487
191 Mits 403 410 20.31% 18.81% 92.64% 391
Nova 5349 4825 15.04% 14.21% 94.48% 4696
Novo 1063 856 15.06% 15.84% 105.16% 927
Nyco 47 36 100.00% 37
Pfiz 7599 7256 14.51% 15.99% 110.20% 8236
Roch 5258 4526 16.07% 15.13% 94.19% 4391
Sano 5565 5080 14.81% 13.40% 90.52% 4736
Schr 2188 1865 19.62% 16.39% 83.58% 1606
Shio 320 276 16.70% 14.30% 85.65% 243
Shir 387 339 100.00% 349
Solv 533 441 15.46% 14.84% 95.95% 436
Tais 244 197 8.49% 9.84% 115.87% 235
Take 1620 1417 15.43% 13.22% 85.66% 1250
Teva 495 369 7.03% 6.32% 89.96% 342
UCB 1024 888 21.83% 14.94% 68.44% 626
Wats 131 125 100.00% 129
Wyet 3109 2749 14.64% 13.69% 93.51% 2648
192 A.5 Conclusion
The historic R&D data has been adjusted to arrive at a figure that best reflects the inputs to the R&D process by adjusting for changes to the policy of R&D expenditure (expressed as a expenditure as a percentage of Revenue) from
2001 onwards and applying an adjustment to the 2005 R&D expenditure and adjusting for inflation. However the historic figure remains highly correlated to the current figure (for 2006) and so the impact on the DEA results may be limited as is the application or removal of input weight restrictions on the relative weight of the current and historic R&D expenditure.
193 B. Staff Data
B.1 Introduction
Some of the DEA models require the numbers of staff as an input. The model input has been calculated by forming the arithmetic means of staff numbers from 2001 to 2005, where that information is available; where not, estimates are made from other sources. The primary data are summarised below.
B.2 Staff Data
The staff data are shown in Table B.1.
Table B.1 Staff Numbers
Staff Staff Staff Staff Staff
Company Staff 2005 2004 2003 2002 2001
Abbt 67155.6 59735 60617 72181 71819 71426
Akzo 64314 61340 61450 64580 67900 66300
Alco 12040 12700 12200 11900 11800 11600
Alle 5270.2 5055 5030 4930 4900 6436
Amg 12320 16500 14400 12900 10118 7682
Astel 11505.4 15000 15000 9062 9278 9187
Astr 60820 63500 64200 62600 59200 54600
Baus 12220 14000 12400 11600 11500 11600
194 Baxt 49780 47000 48000 51300 54600 48000
Baye 108060 93700 91700 115400 122600 116900
Biog 3777.667 3340 4266 3727
BMS 44000 43000 43000 44000 44000 46000
Boeh 33395.8 37406 35529 34221 31843 27980
Ceph 3851.75 3844 3851 3983 3729
Chug 5420.4 5357 5327 5680 5774 4964
CSL 10000
Daic 18605.5 18434 18777
Dain 3016.2 5142 2427 2480 2480 2552
Eisa 7953.8 9081 8295 7700 7433 7260
EliL 43100 42600 44500 45000 42900 40500
Fore 4624.8 5050 5136 4967 4240 3731
Gene 6727.4 9563 7646 6226 5252 4950
Genz 6325 8200 7000 5625 5600 5200
Gile 4000
GSK 102727 100728 100019 100919 104499 107470
John 109240 115600 109900 110600 108300 101800
195 King 2600
Kyow 6429.4 5800 5960 6294 6794 7299
Lund 4618 5155 5223 4534 3560
Merc 68540 61500 62600 63200 77300 78100
Merk 32202.8 29133 28877 34206 34504 34294
Mits 7138.8 5902 5917 6122 8733 9020
Nova 78970 90924 81392 78541 72877 71116
Novo 19038.8 22007 20285 18756 18005 16141
Nyco 12000
Pfiz 106200 106000 115000 122000 98000 90000
Roch 66309 68218 64594 65357 69659 63717
Sano 69942.8 97181 96439 93144 32436 30514
Schr 30780 32600 30500 30500 30500 29800
Shio 6285.2 4997 5522 5589 6149 9169
Shir 4000
Solv 29502 28730 26926 30139 30302 31413
Tais 5141.4 5191 5339 5477 4806 4894
Take 14645.8 15069 14510 14592 14547 14511
196 Teva 11606.6 14698 13813 10960 9576 8986
UCB 9804.2 8525 8598 11559 10326 10013
Wats 5830
Wyet 51713.6 49732 51401 52384 52762 52289
197 C. Power Laws
C.1 Introduction
This appendix summarises the main features of a ‘power law’ and some of the pitfalls in identifying a power law from empirical data. Power laws have recently attracted attention by their apparent ubiquity; however this has now been matched by a critical attitude to their identification. This appendix first defines a power-law distribution and notes alternative distributions which might also give rise to a similar ‘signature’, namely a straight line when the distribution is plotted on a log/log scales. It concludes with a summary of the relevance to the literature of this research and potential future work.
C.2 Definition of a Power Law
The characteristics of a power law are its scale invariance. To illustrate with the simplest expression of a power law:
f(x) = k x a Eq. (C.1)
where x is an independent variable
k is a constant
a is a constant
If there is a scaling transformation given by:
y=c•.x Eq.(C.2)
where c is a constant then:
198 a f(y) = c • f(x) Eq. (C.3) that is the functional form repaints the same with a change in scale. Eq. (C.1) can be written in logarithmic form:
log f(x) = - a. log (x) + log (k) Eq. (C.4)
where x is an independent variable
k is a constant
a is a constant
Eq. (C.4) gives rise to a common means of identification of a power law, a
‘signature’, namely the observance of a straight line on a log/log graph.
Although this is a characteristic of a power law as defined above, two issues arise.
The first is that power laws only apply for a range of variables and it is necessary to establish the range over which the law holds. The second is that other mathematical functions also can show a linear plot on a log/log graph therefore it is necessary to eliminate these possibilities before a power law is confirmed. These issues are now considered further.
C.3 Alternative Distributions
Clauset et al. (2009) considered four discrete and five continuous nonlinear distributions, with the five continuous distributions were a power law; a power law with cut-off, exponential, stretched exponential and log-normal.
After developing statistical testing methods using synthetic data, 24 actual data sets which showed a straight-line log/log plot were tested in order to
199 confirm a power law with empirical data. In only one case, the frequency of words in the English language, could power law be confirmed and all other possibilities excluded. In all but three cases the exponential could be excluded but the log-normal and stretched exponential distributions were plausible alternatives in nine cases. The authors then emphasise the importance of looking at the underlying processes giving rise to the distribution rather than relying on tests alone.
Farmer & Geanakoplos (2008) outlined several mechanisms for generating power laws (as opposed to log-normal which are produced by multiplicative processes) including maximisation of entropy (i.e. randomness), preferential attachment (i.e. a quantity is allocated on the basis of how much is already held) and critical systems (an example of a pile of sand is use to illustrate, where a steady stream of sand will eventually lead to an avalanche of sand).
Mitzenmacher (2001) also considers generative models that produce both power law and log-normal distributions.
C.4 Relevance to this Thesis
A linear plot on a log–log graph has been observed, which could represent either a power law or log-normal behaviour. A plausible explanation for power- law behaviour has been proposed by Park et al. (2010), namely self- organising criticality (with an initial mega-merger leading to a cascade of smaller events); the findings of the thesis are consistent with this hypothesis but are not definitive in establishing a power law. Confirmation of that would require further statistical analysis.
200 D. DEA Model Data
Table D.1 DEA Model Inputs
Symbol for Input x1 x2 x3
Name of Firm Code R&D Expense R&D Expense Staff
US$m, 2006 US$m, Historic Mean Numbers
Abbot Abbt 2255 2208 67156
Laboratories
Akzo Nobell NV Akzo 741 676 64314
Alcon Inc. Alco 512 416 12040
Allergan Inc. Alle 1056 546 5270
Amgen Inc. Amg 3366 2615 12320
Astellas Pharma Astel 1435 1477 11505
Inc.
AstraZeneca Plc Astr 3902 4066 60820
Bausch & Lomb Baus 197 175 12220
Inc.
Baxter Baxt 614 599 49780
International Inc.
Bayer AG Baye 1791 1511 108060
Biogen Idec Inc. Biog 718 747 3778
201 Bristol Myers BMS 3067 2558 44000
Squibb Co.
Boehringer Boeh 1977 1958 33396
Ingelheim
Cephalon Ceph 403 330 3852
Chugai Chug 467 563 5420
Pharmaceutical
CSL Ltd CSL 161 140 10000
Daiichi Sankyo Daii 1459 1308 18606
Co.
Dainippon Dain 350 165 3016
Sumiformo
Eisai Co. Eisa 926 663 7954
Eli Lilly and Co. EliL 3129 2898 43100
Forest Fore 941 282 4625
Pharmaceuticals
Genentech Inc. Gene 1773 1874 6727
Genzyme Corp. Genz 650 578 6325
Gilead Sciences Gile 384 286 4000
Inc.
202 GlaxoSmithKline GSK 6373 5605 102727
Plc.
Johnson & John 7125 5885 109240
Johnson
King King 254 271 2600
Pharmaceuticals
Kyowa Hakko Kyow 268 247 6429
Kogyo
H Lundbeck Lund 329 290 4618
Merck & Co. Merc 4783 3179 68540
Merck KgaA Merk 772 487 32203
Mitsubishi Mits 403 391 7139
Pharma
Novartis Nova 5349 4696 78970
Novo Nordisk As Novo 1063 927 19039
Nycomed Nyco 47 37 12000
Pfizer Inc. Pfiz 7599 8236 106200
Roche Roch 5258 4391 66309
Sanofi Aventis Sano 5565 4736 69943
Group
203 Schering-Plough Scher 2188 1606 30780
Corp.
Shionogi & Co. Shio 320 243 6285
Shire Plc Shir 387 349 4000
Solvay SA Solv 533 436 29502
Taisho Tais 244 235 5141
Pharmaceutical
Takeda Take 1620 1250 14646
Pharmaceutical
Teva Teva 495 342 11607
Pharmaceutical
UCB SA UCB 1024 626 9804
Watson Wats 131 129 5830
Pharmaceutical
Wyeth Wyet 3109 2648 51714
204 Table D.2 DEA Model Compounds (Comp.) Output Data
Symbol y5 y4 y3 y2 y1
Phase Awaiting Phase 3 Phase 2 Phase 1 Preclinical
Approval Comp. Comp. Comp. Comp.
Abbt 3 9 7 11 10
Akzo 2 4 4 9 7
Alco 4 3 2 0 0
Alle 4 5 6 0 2
Amg 2 8 11 14 1
Astel 15 6 15 1 3
Astr 5 14 14 25 38
Baus 0 1 1 0 1
Baxt 0 1 1 0 2
Baye 9 15 16 14 2
Biog 1 8 11 1 10
BMS 9 7 4 8 1
Boeh 2 2 3 0 0
Ceph 1 5 5 0 4
205 Chug 7 5 10 5 0
CSL 6 4 1 1 2
Daii 4 10 10 13 1
Dain 4 1 11 1 0
Eisa 8 7 7 8 2
EliL 6 11 21 12 8
Fore 3 4 3 0 3
Gene 1 13 12 15 3
Genz 6 8 8 8 8
Gile 2 3 1 2 3
GSK 23 24 30 40 1
John 8 23 8 8 1
King 1 4 3 0 0
Kyow 2 2 3 3 1
Lund 0 3 2 4 0
Merc 11 7 17 30 2
Merk 3 7 14 8 8
Mits 4 7 9 0 0
206 Nova 15 30 28 26 9
Novo 2 5 5 6 0
Nyco 1 5 4 1 4
Pfiz 5 11 52 42 5
Roch 7 18 22 30 6
Sano 14 24 35 34 39
Scher 10 13 13 2 2
Shio 3 1 7 4 0
Shir 4 1 2 1 4
Solv 8 10 7 4 1
Tais 0 0 8 3 0
Take 7 14 12 5 1
Teva 0 5 6 0 4
UCB 4 7 4 0 1
Wats 2 2 0 1 0
Wyet 7 10 13 2 1
207 Table D.3 DEA Model Clinical Trials Output Data
Sym- y5 y4 y3 y2 y1 Ratio Trials to Compounds bol
Phase Await Ph. 3 Ph. 2 Ph. 1 Precl AA P3 P2 P1 PC
(Ph.) .Appr (P3) (P2) (P1) inic
(AA) (PC)
Abbt 8 10 10 15 11 2.67 1.11 1.43 1.36 1.10
Akzo 3 8 6 9 9 1.50 2.00 1.50 1.00 1.29
Alco 4 6 2 0 0 1.00 2.00 1.00
Alle 4 10 8 0 2 1.00 2.00 1.33 1.00
Amg 3 16 12 15 1 1.50 2.00 1.09 1.07 1.00
Astel 22 12 17 1 4 1.47 2.00 1.13 1.00 1.33
Astr 8 42 33 53 45 1.60 3.00 2.36 2.12 1.18
Baus 0 1 1 0 1 1.00 1.00 1.00
Baxt 0 1 1 1 2 1.00 1.00 1.00
Baye 11 24 32 15 2 1.22 1.60 2.00 1.07 1.00
Biog 3 24 18 1 11 3.00 3.00 1.64 1.00 1.10
BMS 11 14 8 9 1 1.22 2.00 2.00 1.13 1.00
Boeh 2 4 3 0 0 1.00 2.00 1.00
208 Ceph 1 8 5 0 4 1.00 1.60 1.00 1.00
Chug 9 7 17 6 0 1.29 1.40 1.70 1.20
CSL 6 5 1 1 3 1.00 1.25 1.00 1.00 1.50
Daii 5 12 22 26 2 1.25 1.20 2.20 2.00 2.00
Dain 4 1 19 1 0 1.00 1.00 1.73 1.00
Eisa 10 12 24 10 2 1.25 1.71 3.43 1.25 1.00
EliL 7 19 28 12 12 1.17 1.73 1.33 1.00 1.50
Fore 4 5 4 0 4 1.33 1.25 1.33 1.33
Gene 1 46 21 20 4 1.00 3.54 1.75 1.33 1.33
Genz 6 12 10 10 10 1.00 1.50 1.25 1.25 1.25
Gile 4 3 2 3 4 2.00 1.00 2.00 1.50 1.33
GSK 27 35 87 49 1 1.17 1.46 2.90 1.23 1.00
John 15 48 9 9 1 1.88 2.09 1.13 1.13 1.00
King 1 4 7 0 0 1.00 1.00 2.33
Kyow 2 2 3 4 1 1.00 1.00 1.00 1.33 1.00
Lund 0 3 2 4 0 1.00 1.00 1.00
Merc 14 12 29 33 2 1.27 1.71 1.71 1.10 1.00
Merk 3 10 34 19 10 1.00 1.43 2.43 2.38 1.25
209 Mits 5 11 11 0 0 1.25 1.57 1.22
Nova 21 42 38 30 10 1.40 1.40 1.36 1.15 1.11
Novo 7 7 9 11 0 3.50 1.40 1.80 1.83
Nyco 1 7 7 1 8 1.00 1.40 1.75 1.00 2.00
Pfiz 10 19 64 42 5 2.00 1.73 1.23 1.00 1.00
Roch 14 60 38 35 6 2.00 3.33 1.73 1.17 1.00
Sano 14 37 53 42 52 1.00 1.54 1.51 1.24 1.33
Scher 10 26 18 2 2 1.00 2.00 1.38 1.00 1.00
Shio 3 1 7 4 0 1.00 1.00 1.00 1.00
Shir 4 2 2 1 4 1.00 2.00 1.00 1.00 1.00
Solv 11 11 9 4 1 1.38 1.10 1.29 1.00 1.00
Tais 0 0 14 5 0 1.75 1.67
Take 8 23 29 10 1 1.14 1.64 2.42 2.00 1.00
Teva 0 6 8 0 4 1.20 1.33 1.00
UCB 6 13 6 0 1 1.50 1.86 1.50 1.00
Wats 2 2 0 1 0 1.00 1.00 1.00
Wyet 10 16 19 2 1 1.43 1.60 1.46 1.00 1.00
210 Table D.4 DEA Efficiency Scores for Compounds as Outputs
Symbol ηk θk ek ln(ek) ln ((x1k +x2k)/2)
Firm VRS CRS Scale Log Log Avg.
Eff. Eff. Eff. Scale Eff. R&D
Abbt 53.3% 5.6% 10.5% -2.25494 7.711
Akzo 63.5% 11.4% 17.9% -1.7209 6.563
Alco 50.0% 22.6% 45.2% -0.79386 6.140
Alle 53.1% 15.8% 29.7% -1.21273 6.686
Amg 47.5% 4.5% 9.4% -2.36105 8.003
Astel 100.0% 27.9% 27.9% -1.27773 7.284
Astr 83.2% 7.1% 8.5% -2.46574 8.290
Baus 14.6% 4.6% 31.5% -1.15603 5.226
Baxt 10.7% 1.9% 17.9% -1.72307 6.408
Baye 100.0% 17.9% 17.9% -1.72284 7.409
Biog 78.5% 12.4% 15.8% -1.84477 6.596
BMS 55.1% 8.7% 15.8% -1.84215 7.942
Boeh 16.7% 3.0% 17.9% -1.72215 7.585
Ceph 55.9% 12.7% 22.7% -1.48295 5.903
211 Chug 93.3% 41.2% 44.2% -0.81673 6.244
CSL 100.0% 100.0% 100.0% 0 5.014
Daii 77.1% 11.4% 14.8% -1.91103 7.233
Dain 97.2% 50.7% 52.2% -0.65005 5.550
Eisa 90.0% 29.4% 32.7% -1.1189 6.678
EliL 71.9% 7.5% 10.5% -2.25524 8.011
Fore 40.8% 16.4% 40.1% -0.91263 6.416
Gene 74.2% 9.0% 12.1% -2.10854 7.508
Genz 100.0% 30.4% 30.4% -1.19204 6.420
Gile 43.0% 17.9% 41.5% -0.87945 5.815
GSK 100.0% 12.0% 12.0% -2.12224 8.698
John 66.0% 4.2% 6.3% -2.7648 8.780
King 46.8% 15.2% 32.5% -1.12311 5.570
Kyow 38.6% 14.5% 37.6% -0.97846 5.552
Lund 41.4% 11.3% 27.3% -1.29768 5.734
Merc 75.8% 9.3% 12.3% -2.09723 8.289
Merk 100.0% 20.5% 20.5% -1.58451 6.445
Mits 92.3% 32.7% 35.5% -1.0367 5.984
212 Nova 100.0% 10.2% 10.2% -2.28306 8.522
Novo 45.2% 7.8% 17.2% -1.75763 6.903
Nyco 100.0% 100.0% 100.0% 0 3.739
Pfiz 94.9% 5.6% 5.9% -2.83154 8.977
Roch 75.6% 6.2% 8.1% -2.50842 8.481
Sano 100.0% 9.8% 9.8% -2.31827 8.547
Scher 90.6% 15.5% 17.1% -1.76349 7.548
Shio 72.8% 33.2% 45.7% -0.78342 5.641
Shir 54.3% 28.4% 52.4% -0.64683 5.908
Solv 100.0% 46.8% 46.8% -0.75987 6.183
Tais 58.6% 18.1% 30.9% -1.1732 5.479
Take 91.2% 15.6% 17.1% -1.76728 7.269
Teva 52.0% 11.3% 21.7% -1.52662 6.037
UCB 55.7% 15.0% 27.0% -1.31015 6.715
Wats 43.8% 41.5% 94.9% -0.05219 4.867
Wyet 58.2% 7.4% 12.7% -2.06679 7.965
213 Table D.5 DEA Efficiency Scores for Trials as Outputs
Symbol ηk θk ek ln(ek) ln ((x1k +x2k)/2)
Firm VRS CRS Scale Log Log Avg.
Eff. Eff. Eff. Scale Eff. R&D
Abbt 53.4% 21.0% 39.4% -0.93182 7.711
Akzo 56.2% 22.1% 39.4% -0.93184 6.563
Alco 54.8% 48.5% 88.6% -0.12149 6.140
Alle 72.5% 72.3% 99.8% -0.00152 6.686
Amg 52.5% 30.7% 58.4% -0.53722 8.003
Astel 100.0% 100.0% 100.0% 0 7.284
Astr 100.0% 47.7% 47.7% -0.7397 8.290
Baus 9.2% 8.7% 94.3% -0.0584 5.226
Baxt 7.8% 3.9% 49.4% -0.70459 6.408
Baye 100.0% 35.4% 35.4% -1.03815 7.409
Biog 100.0% 100.0% 100.0% 0 6.596
BMS 53.9% 25.8% 47.9% -0.73658 7.942
Boeh 14.1% 7.5% 53.1% -0.63236 7.585
Ceph 80.8% 67.7% 83.8% -0.17712 5.903
214 Chug 100.0% 100.0% 100.0% 0 6.244
CSL 100.0% 100.0% 100.0% 0 5.014
Daii 84.5% 55.3% 65.5% -0.42351 7.233
Dain 100.0% 100.0% 100.0% 0 5.550
Eisa 100.0% 100.0% 100.0% 0 6.678
EliL 61.8% 27.7% 44.8% -0.80197 8.011
Fore 86.3% 85.7% 99.3% -0.0068 6.416
Gene 100.0% 100.0% 100.0% 0 7.508
Genz 98.0% 91.3% 93.1% -0.07112 6.420
Gile 77.6% 76.8% 99.0% -0.01027 5.815
GSK 100.0% 36.8% 36.8% -0.99975 8.698
John 88.4% 23.9% 27.0% -1.30784 8.780
King 100.0% 68.1% 68.1% -0.38465 5.570
Kyow 47.3% 42.5% 89.9% -0.10641 5.552
Lund 35.6% 33.3% 93.4% -0.06852 5.734
Merc 70.7% 23.3% 32.9% -1.11208 8.289
Merk 100.0% 78.5% 78.5% -0.2424 6.445
Mits 99.0% 98.0% 98.9% -0.01093 5.984
215 Nova 99.1% 33.9% 34.2% -1.07342 8.522
Novo 58.8% 40.4% 68.7% -0.37484 6.903
Nyco 100.0% 100.0% 100.0% 0 3.739
Pfiz 70.4% 21.2% 30.1% -1.20163 8.977
Roch 100.0% 41.6% 41.6% -0.87618 8.481
Sano 100.0% 40.6% 40.6% -0.90077 8.547
Scher 91.5% 50.0% 54.6% -0.6047 7.548
Shio 57.9% 56.9% 98.2% -0.01797 5.641
Shir 68.9% 64.5% 93.7% -0.06514 5.908
Solv 100.0% 71.0% 71.0% -0.34217 6.183
Tais 100.0% 85.8% 85.8% -0.15332 5.479
Take 100.0% 66.6% 66.6% -0.40593 7.269
Teva 44.5% 36.4% 81.9% -0.19981 6.037
UCB 82.1% 75.3% 91.8% -0.08594 6.715
Wats 100.0% 52.9% 52.9% -0.63668 4.867
Wyet 55.9% 21.2% 38.0% -0.96742 7.965
216 Table D.6 DEA Efficiency Scores for VRS for Alternative Input Assumptions
ηk ηk
Firm Input: Input:
R&D Only R&D plus Staff
Abbt 53.3% 53.3%
Akzo 63.5% 63.5%
Alco 50.0% 53.0%
Alle 53.1% 76.1%
Amg 47.5% 75.3%
Astel 100.0% 100.0%
Astr 83.2% 83.2%
Baus 14.6% 14.8%
Baxt 10.7% 10.7%
Baye 100.0% 100.0%
Biog 78.5% 100.0%
BMS 55.1% 55.1%
Boeh 16.7% 16.9%
Ceph 55.9% 77.8%
Chug 93.3% 100.0%
217 CSL 100.0% 100.0%
Daii 77.1% 86.0%
Dain 97.2% 100.0%
Eisa 90.0% 99.2%
EliL 71.9% 74.0%
Fore 40.8% 68.9%
Gene 74.2% 100.0%
Genz 100.0% 100.0%
Gile 43.0% 68.0%
GSK 100.0% 100.0%
John 66.0% 66.0%
King 46.8% 100.0%
Kyow 38.6% 57.8%
Lund 41.4% 47.6%
Merc 75.8% 77.5%
Merk 100.0% 100.0%
Mits 92.3% 100.0%
Nova 100.0% 100.0%
218 Novo 45.2% 48.6%
Nyco 100.0% 100.0%
Pfiz 94.9% 94.9%
Roch 75.6% 76.0%
Sano 100.0% 100.0%
Scher 90.6% 93.1%
Shio 72.8% 80.9%
Shir 54.3% 76.5%
Solv 100.0% 100.0%
Tais 58.6% 68.6%
Take 91.2% 100.0%
Teva 52.0% 55.8%
UCB 55.7% 69.6%
Wats 43.8% 100.0%
Wyet 58.2% 59.1%
219 Table D.7 ROS, ROA and SOA
Firm ROS (%) ROA (%) SOA
Abbt 7.64 6.17 0.81
AkzN 8.39 11.42 1.36
Alcn 27.53 26.26 0.95
Allg -4.16 -2.08 0.50
Amgn 20.68 9.62 0.47
Astel 11.79 8.50 0.72
AstrZ 22.83 23.04 1.01
BausL 0.65 1.85 2.85
Baxt 13.46 11.38 0.85
Bayr 6.06 6.29 1.04
Biog 12.20 2.50 0.20
Boeh 15.67 14.54 0.93
BrMS 8.85 7.74 0.87
Ceph 8.20 4.76 0.58
Chug 11.78 8.60 0.73
CSL 16.30 18.40 1.13
220 Daic 9.47 6.89 0.73
Dain 6.26 5.20 0.83
Eisa 10.55 9.32 0.88
EliL 16.97 12.11 0.71
Fore 25.36 20.82 0.82
Gene 22.76 16.21 0.71
Genz -0.53 -0.10 0.18
Gild -39.30 -29.10 0.74
GSK 23.20 23.28 1.00
Hlun 12.00 10.11 0.84
John 20.73 17.74 0.86
King 14.53 8.68 0.60
Kyow 4.60 4.34 0.94
Merc 19.59 10.49 0.54
MerK 15.71 13.83 0.88
Mits 9.02 5.64 0.63
Nova 19.91 12.40 0.62
Novo 16.65 15.88 0.95
221 Nyco -9.59 -0.91 0.09
Pfiz 39.98 17.01 0.43
RocH 18.74 11.70 0.62
Sano 14.12 5.47 0.39
Schr 10.79 7.96 0.74
Shio 11.58 5.53 0.48
Shir 15.50 8.30 0.54
Solv 8.42 8.25 0.98
Tais 13.22 5.66 0.43
Take 25.84 11.27 0.44
Teva 6.49 4.53 0.70
UCB 16.77 5.56 0.33
Wats -22.50 -0.12 0.01
Wyet 20.62 13.09 0.63
222 Table D.8 SDV, Cost of Sales and NDV Data for Firms
DMU Ak Bk Ck Dk ak bk ck
Aggreg. X- X- Cost of Aggreg. X- X-
SDV border product Sales NDV border product
SDV SDV NDV NDV
Abbt 11938 7674 3027 20759 0.58 0.37 0.15
AkzN 4452 4452 711 3245 1.37 1.37 0.22
Alcn 0 0 0 3549 0.00 0.00 0.00
Allg 490 0 230 3190 0.15 0.00 0.07
Amgn 18276 138 0 11318 1.61 0.01 0.00
Astl 0 0 0 6728 0.00 0.00 0.00
Astz 39021 39021 644 20412 1.91 1.91 0.03
Baus 1237 427 1009 2277 0.54 0.19 0.44
Baxt 2652 1055 801 8981 0.30 0.12 0.09
Bayr 14530 14530 10469 37750 0.38 0.38 0.28
Biog 0 0 0 2465 0.00 0.00 0.00
Boeh 0 0 0 11121 0.00 0.00 0.00
BrMS 8212 150 0 16329 0.50 0.01 0.00
Ceph 1998 810 0 1619 1.23 0.50 0.00
Chug 2590 0 0 2459 1.05 0.00 0.00
223 CSL 1669 1669 0 2334 0.72 0.72 0.00
Daii 6290 0 0 6487 0.97 0.00 0.00
Dain 2224 0 0 1570 1.42 0.00 0.00
Eisi 265 265 0 4979 0.05 0.05 0.00
EliL 4381 0 0 13028 0.34 0.00 0.00
Fors 0 0 0 2988 0.00 0.00 0.00
Gene 408 0 0 7171 0.06 0.00 0.00
Genz 4710 107 250 3204 1.47 0.03 0.08
Gild 1396 0 0 4216 0.33 0.00 0.00
GSK 102218 10035 1453 32678 3.13 0.31 0.04
Hlun 236 0 236 1366 0.17 0.00 0.17
John 27524 489 5083 42271 0.65 0.01 0.12
King 5741 637 235 1700 3.38 0.37 0.14
Kyow 0 0 0 1590 0.00 0.00 0.00
Merc 6567 0 0 18202 0.36 0.00 0.00
MerK 2551 0 0 3917 0.65 0.00 0.00
Mits 0 0 0 1737 0.00 0.00 0.00
Nova 14629 14629 1859 29818 0.49 0.49 0.06
224 Novo 0 0 0 5434 0.00 0.00 0.00
Nyco 0 0 0 4369 0.00 0.00 0.00
Pfiz 163448 7667 356 29034 5.63 0.26 0.01
RocH 25129 17138 2430 26229 0.96 0.65 0.09
Sano 71858 0 0 30612 2.35 0.00 0.00
Sche 1572 1167 405 9537 0.16 0.12 0.04
Shio 120 120 0 1481 0.08 0.08 0.00
Shir 6528 6528 0 1519 4.30 4.30 0.00
Slvy 112 0 0 2242 0.05 0.00 0.00
Tais 0 0 0 1937 0.00 0.00 0.00
Take 270 270 0 7414 0.04 0.04 0.00
Teva 3988 0 0 7862 0.51 0.00 0.00
UCBs 2973 2973 0 3934 0.76 0.76 0.00
Wats 2259 0 0 2424 0.93 0.00 0.00
Wyet 0 0 0 16154 0.00 0.00 0.00
225 Table D.9 NDV and Pure Technical Efficiency (Base Model)
Firm VRS ak ak bk bk ck ck
Efficiency θk
Abbt 53.3% 0.58 0.37 0.15
AkzN 63.5% 1.37 1.37 0.22
Alcn 50.0% 0.00 0.00 0.00
Allg 53.1% 0.15 0.00 0.07
Amgn 47.5% 1.61 0.01 0.00
Astel 100.0% 0.00 0.00 0.00
AstrZ 83.2% 1.91 1.91 0.03
BausL 14.6% 0.54 0.19 0.44
Baxt 10.7% 0.30 0.12 0.09
Bayr 100.0% 0.38 0.38 0.28
Biog 78.5% 0.00 0.00 0.00
Boeh 55.1% 0.00 0.00 0.00
BrMS 16.7% 0.50 0.01 0.00
Ceph 55.9% 1.23 0.50 0.00
Chug 93.3% 1.05 0.00 0.00
CSL 100.0% 0.72 0.72 0.00
Daic 77.1% 0.97 0.00 0.00
Dain 97.2% 1.42 0.00 0.00
Eisa 90.0% 0.05 0.05 0.00
EliL 71.9% 0.34 0.00 0.00
Fore 40.8% 0.00 0.00 0.00
Gene 74.2% 0.06 0.00 0.00
Genz 100.0% 1.47 0.03 0.08
226 Gild 43.0% 0.33 0.00 0.00
GSK 100.0% 3.13 0.31 0.04
Hlun 66.0% 0.17 0.00 0.17
John 46.8% 0.65 0.01 0.12
King 38.6% 3.38 0.37 0.14
Kyow 41.4% 0.00 0.00 0.00
Merc 75.8% 0.36 0.00 0.00
MerK 100.0% 0.65 0.00 0.00
Mits 92.3% 0.00 0.00 0.00
Nova 100.0% 0.49 0.49 0.06
Novo 45.2% 0.00 0.00 0.00
Nyco 100.0% 0.00 0.00 0.00
Pfiz 94.9% 5.63 0.26 0.01
RocH 75.6% 0.96 0.65 0.09
Sano 100.0% 2.35 0.00 0.00
Schr 90.6% 0.16 0.12 0.04
Shio 72.8% 0.08 0.08 0.00
Shir 54.3% 4.30 4.30 0.00
Solv 100.0% 0.05 0.00 0.00
Tais 58.6% 0.00 0.00 0.00
Take 91.2% 0.04 0.04 0.00
Teva 52.0% 0.51 0.00 0.00
UCB 55.7% 0.76 0.76 0.00
Wats 43.8% 0.93 0.00 0.00
Wyet 58.2% 0.00 0.00 0.00
Mean 0.74 0.91 0.33 0.21 0.06 0.03
227 Table D.10 NDV and Pure Technical Efficiency (Staff Input)
Firm VRS ak ak bk bk ck ck
Efficiency θk
Abbt 53.3% 0.58 0.37 0.15
AkzN 63.5% 1.37 1.37 0.22
Alcn 53.0% 0.00 0.00 0.00
Allg 76.1% 0.15 0.00 0.07
Amgn 75.3% 1.61 0.01 0.00
Astel 100.0% 0.00 0.00 0.00
AstrZ 83.2% 1.91 1.91 0.03
BausL 14.8% 0.54 0.19 0.44
Baxt 10.7% 0.30 0.12 0.09
Bayr 100.0% 0.38 0.38 0.28
Biog 100.0% 0.00 0.00 0.00
Boeh 55.1% 0.00 0.00 0.00
BrMS 16.9% 0.50 0.01 0.00
Ceph 77.8% 1.23 0.50 0.00
Chug 100.0% 1.05 0.00 0.00
CSL 100.0% 0.72 0.72 0.00
Daic 86.0% 0.97 0.00 0.00
Dain 100.0% 1.42 0.00 0.00
Eisa 99.2% 0.05 0.05 0.00
EliL 74.0% 0.34 0.00 0.00
Fore 68.9% 0.00 0.00 0.00
Gene 100.0% 0.06 0.00 0.00
Genz 100.0% 1.47 0.03 0.08
Gild 68.0% 0.33 0.00 0.00
228 GSK 100.0% 3.13 0.31 0.04
Hlun 66.0% 0.17 0.00 0.17
John 100.0% 0.65 0.01 0.12
King 57.8% 3.38 0.37 0.14
Kyow 47.6% 0.00 0.00 0.00
Merc 77.5% 0.36 0.00 0.00
MerK 100.0% 0.65 0.00 0.00
Mits 100.0% 0.00 0.00 0.00
Nova 100.0% 0.49 0.49 0.06
Novo 48.6% 0.00 0.00 0.00
Nyco 100.0% 0.00 0.00 0.00
Pfiz 94.9% 5.63 0.26 0.01
RocH 76.0% 0.96 0.65 0.09
Sano 100.0% 2.35 0.00 0.00
Schr 93.1% 0.16 0.12 0.04
Shio 80.9% 0.08 0.08 0.00
Shir 76.5% 4.30 4.30 0.00
Solv 100.0% 0.05 0.00 0.00
Tais 68.6% 0.00 0.00 0.00
Take 100.0% 0.04 0.04 0.00
Teva 55.8% 0.51 0.00 0.00
UCB 69.6% 0.76 0.76 0.00
Wats 100.0% 0.93 0.00 0.00
Wyet 59.1% 0.00 0.00 0.00
Mean 0.72 0.92 0.36 0.18 0.06 0.03
229 Table D.11 SDV and Pure Technical Efficiency (Base Model)
Firm VRS Ak Ak Bk Bk Ck Ck
Efficiency θk
Abbt 53.3% 11938 7674 3027
AkzN 63.5% 4452 4452 711
Alcn 53.0% 0 0 0
Allg 76.1% 490 0 230
Amgn 75.3% 18276 138 0
Astel 100.0% 0 0 0
AstrZ 83.2% 39021 39021 644
BausL 14.8% 1237 427 1009
Baxt 10.7% 2652 1055 801
Bayr 100.0% 14530 14530 10469
Biog 100.0% 0 0 0
Boeh 55.1% 0 0 0
BrMS 16.9% 8212 150 0
Ceph 77.8% 1998 810 0
Chug 100.0% 2590 0 0
CSL 100.0% 1669 1669 0
230 Daic 86.0% 6290 0 0
Dain 100.0% 2224 0 0
Eisa 99.2% 265 265 0
EliL 74.0% 4381 0 0
Fore 68.9% 0 0 0
Gene 100.0% 408 0 0
Genz 100.0% 4710 107 250
Gild 68.0% 1396 0 0
GSK 100.0% 102218 10035 1453
Hlun 66.0% 236 0 236
John 100.0% 27524 489 5083
King 57.8% 5741 637 235
Kyow 47.6% 0 0 0
Merc 77.5% 6567 0 0
MerK 100.0% 2551 0 0
Mits 100.0% 0 0 0
Nova 100.0% 14629 14629 1859
Novo 48.6% 0 0 0
231 Nyco 100.0% 0 0 0
Pfiz 94.9% 163448 7667 356
RocH 76.0% 25129 17138 2430
Sano 100.0% 71858 0 0
Schr 93.1% 1572 1167 405
Shio 80.9% 120 120 0
Shir 76.5% 6528 6528 0
Solv 100.0% 112 0 0
Tais 68.6% 0 0 0
Take 100.0% 270 270 0
Teva 55.8% 3988 0 0
UCB 69.6% 2973 2973 0
Wats 100.0% 2259 0 0
Wyet 59.1% 0 0 0
Mean 4345 19174 1056 4442 472 744
232 Table D.12 Association of NDV and ROS
Firm ROS a’k a’k b’k b’k c’k c’k
rk
Abbt 7.64 0.58 0.37 0.15
AkzN 8.39 1.37 1.37 0.22
Alcn 27.53 0.00 0.00 0.00
Allg -4.16 0.15 0.00 0.07
Amgn 20.68 1.61 0.01 0.00
Astel 11.79 0.00 0.00 0.00
AstrZ 22.83 1.91 1.91 0.03
BausL 0.65 0.54 0.19 0.44
Baxt 13.46 0.30 0.12 0.09
Bayr 6.06 0.38 0.38 0.28
Biog 12.20 0.00 0.00 0.00
Boeh 15.67 0.00 0.00 0.00
BrMS 8.85 0.50 0.01 0.00
Ceph 8.20 1.23 0.50 0.00
Chug 11.78 1.05 0.00 0.00
CSL 16.30 0.72 0.72 0.00
Daic 9.47 0.97 0.00 0.00
233 Dain 6.26 1.42 0.00 0.00
Eisa 10.55 0.05 0.05 0.00
EliL 16.97 0.34 0.00 0.00
Fore 25.36 0.00 0.00 0.00
Gene 22.76 0.06 0.00 0.00
Genz -0.53 1.47 0.03 0.08
Gild -39.30 0.33 0.00 0.00
GSK 23.20 3.13 0.31 0.04
Hlun 12.00 0.17 0.00 0.17
John 20.73 0.65 0.01 0.12
King 14.53 3.38 0.37 0.14
Kyow 4.60 0.00 0.00 0.00
Merc 19.59 0.36 0.00 0.00
MerK 15.71 0.65 0.00 0.00
Mits 9.02 0.00 0.00 0.00
Nova 19.91 0.49 0.49 0.06
Novo 16.65 0.00 0.00 0.00
Nyco -9.59 0.00 0.00 0.00
Pfiz 39.98 5.63 0.26 0.01
234 RocH 18.74 0.96 0.65 0.09
Sano 14.12 2.35 0.00 0.00
Schr 10.79 0.16 0.12 0.04
Shio 11.58 0.08 0.08 0.00
Shir 15.50 4.30 4.30 0.00
Solv 8.42 0.05 0.00 0.00
Tais 13.22 0.00 0.00 0.00
Take 25.84 0.04 0.04 0.00
Teva 6.49 0.51 0.00 0.00
UCB 16.77 0.76 0.76 0.00
Wats -22.50 0.93 0.00 0.00
Wyet 20.62 0.00 0.00 0.00
Mean 0.50 1.15 0.13 0.41 0.06 0.02
235 Table D.13 Association of NDV and ROA
Firm ROA a’k a’k b’k b’k c’k c’k
rk
Abbt 6.17 0.58 0.37 0.15
AkzN 11.42 1.37 1.37 0.22
Alcn 26.26 0.00 0.00 0.00
Allg -2.08 0.15 0.00 0.07
Amgn 9.62 1.61 0.01 0.00
Astel 8.50 0.00 0.00 0.00
AstrZ 23.04 1.91 1.91 0.03
BausL 1.85 0.54 0.19 0.44
Baxt 11.38 0.30 0.12 0.09
Bayr 6.29 0.38 0.38 0.28
Biog 2.50 0.00 0.00 0.00
Boeh 14.54 0.00 0.00 0.00
BrMS 7.74 0.50 0.01 0.00
Ceph 4.76 1.23 0.50 0.00
Chug 8.60 1.05 0.00 0.00
CSL 18.40 0.72 0.72 0.00
Daic 6.89 0.97 0.00 0.00
236 Dain 5.20 1.42 0.00 0.00
Eisa 9.32 0.05 0.05 0.00
EliL 12.11 0.34 0.00 0.00
Fore 20.82 0.00 0.00 0.00
Gene 16.21 0.06 0.00 0.00
Genz -0.10 1.47 0.03 0.08
Gild -29.10 0.33 0.00 0.00
GSK 23.28 3.13 0.31 0.04
Hlun 10.11 0.17 0.00 0.17
John 17.74 0.65 0.01 0.12
King 8.68 3.38 0.37 0.14
Kyow 4.34 0.00 0.00 0.00
Merc 10.49 0.36 0.00 0.00
MerK 13.83 0.65 0.00 0.00
Mits 5.64 0.00 0.00 0.00
Nova 12.40 0.49 0.49 0.06
Novo 15.88 0.00 0.00 0.00
Nyco -0.91 0.00 0.00 0.00
Pfiz 17.01 5.63 0.26 0.01
237 RocH 11.70 0.96 0.65 0.09
Sano 5.47 2.35 0.00 0.00
Schr 7.96 0.16 0.12 0.04
Shio 5.53 0.08 0.08 0.00
Shir 8.30 4.30 4.30 0.00
Solv 8.25 0.05 0.00 0.00
Tais 5.66 0.00 0.00 0.00
Take 11.27 0.04 0.04 0.00
Teva 4.53 0.51 0.00 0.00
UCB 5.56 0.76 0.76 0.00
Wats -0.12 0.93 0.00 0.00
Wyet 13.09 0.00 0.00 0.00
Mean 0.70 0.95 0.28 0.26 0.044 0.041
238 Table D.14 Association of Acquisition History and SOA
Firm SOA a’k a’k b’k b’k c’k c’k
rk
Abbt 0.81 0.58 0.37 0.15
AkzN 1.36 1.37 1.37 0.22
Alcn 0.95 0.00 0.00 0.00
Allg 0.50 0.15 0.00 0.07
Amgn 0.47 1.61 0.01 0.00
Astel 0.72 0.00 0.00 0.00
AstrZ 1.01 1.91 1.91 0.03
BausL 2.85 0.54 0.19 0.44
Baxt 0.85 0.30 0.12 0.09
Bayr 1.04 0.38 0.38 0.28
Biog 0.20 0.00 0.00 0.00
Boeh 0.93 0.00 0.00 0.00
BrMS 0.87 0.50 0.01 0.00
Ceph 0.58 1.23 0.50 0.00
Chug 0.73 1.05 0.00 0.00
CSL 1.13 0.72 0.72 0.00
Daic 0.73 0.97 0.00 0.00
239 Dain 0.83 1.42 0.00 0.00
Eisa 0.88 0.05 0.05 0.00
EliL 0.71 0.34 0.00 0.00
Fore 0.82 0.00 0.00 0.00
Gene 0.71 0.06 0.00 0.00
Genz 0.18 1.47 0.03 0.08
Gild 0.74 0.33 0.00 0.00
GSK 1.00 3.13 0.31 0.04
Hlun 0.84 0.17 0.00 0.17
John 0.86 0.65 0.01 0.12
King 0.60 3.38 0.37 0.14
Kyow 0.94 0.00 0.00 0.00
Merc 0.54 0.36 0.00 0.00
MerK 0.88 0.65 0.00 0.00
Mits 0.63 0.00 0.00 0.00
Nova 0.62 0.49 0.49 0.06
Novo 0.95 0.00 0.00 0.00
Nyco 0.09 0.00 0.00 0.00
Pfiz 0.43 5.63 0.26 0.01
240 RocH 0.62 0.96 0.65 0.09
Sano 0.39 2.35 0.00 0.00
Schr 0.74 0.16 0.12 0.04
Shio 0.48 0.08 0.08 0.00
Shir 0.54 4.30 4.30 0.00
Solv 0.98 0.05 0.00 0.00
Tais 0.43 0.00 0.00 0.00
Take 0.44 0.04 0.04 0.00
Teva 0.70 0.51 0.00 0.00
UCB 0.33 0.76 0.76 0.00
Wats 0.01 0.93 0.00 0.00
Wyet 0.63 0.00 0.00 0.00
Mean 1.03 0.62 0.31 0.23 0.02 0.07
241 E. Acquisition Data
E.1 Introduction
The Thomson Banker database was the source of acquisition data. The search criteria used are explained below and an annual analysis of the acquisitions that arose is then presented. There is then a record of each acquisition, characterised by date, value, and the sector and nation of the acquirer and acquired company, sorted by eventual 'surviving' parent.
E.2 Search Criteria
The search criteria are shown in Table E.1
242 Table E.1 Database Search Criteria
Request Operator Description Hits
Acquirer NAIC Include Medicinal and Botanical 13500
(Code) Manufacturing
Pharmaceutical Preparation
Manufacturing
In-Vitro Diagnostic Substance
Manufacturing
Biological Product (except
Diagnostic) Manufacturing
Acquirer Ultimate Include Medicinal and Botanical 8906
Parent Primary Manufacturing
NAIC (Code) Pharmaceutical Preparation
Manufacturing
In-Vitro Diagnostic Substance
Manufacturing
Biological Product (except
Diagnostic) Manufacturing
Logical Set Request # 2 13500
UNION Request # 3
Date Effective/ Between 01/01/1993 to 31/12/2006 5816
Unconditional
243 Ranking Value inc. Between 100 to HI 699
Net Debt of Target
($Mil)
Per cent of Shares Between 51 to HI 591
Owned after
Transaction
Four sectors have been chosen for the analysis namely Medicinal and
Botanical Manufacturing, Pharmaceutical Preparation Manufacturing, In-Vitro
Diagnostic Substance Manufacturing, and Biological Product (except
Diagnostic) Manufacturing. The acquisitions of interest are when the acquiring firm or the ultimate acquiring firm fall within these sectors. There are 13,500 such deals.
The period of the analysis was from the start of 1993 to the end of 2006. This reduces the numbers of deals within the criteria to 5,818. A threshold of the deal value being above $100million was also set, reducing the number of deals to 689. Finally, only deals leading to majority control were considered, reducing the deal total to 591.
E.3 Annual Analysis
A breakdown of deals by year is given in Table E.2.
244 Table E.2 Date Analysis
Date Value ($Mil) Share (%) No. Deals
1992 1,022.75 0.1 3
1993 8,962.57 1.0 13
1994 36,332.34 3.9 24
1995 31,792.50 3.4 24
1996 13,501.48 1.5 34
1997 25,894.22 2.8 36
1998 62,700.57 6.7 40
1999 145,714.96 15.7 42
2000 116,194.32 12.5 49
2001 65,090.71 7.0 46
2002 72,760.70 7.8 37
2003 33,593.12 3.6 48
2004 105,818.15 11.4 51
2005 137,117.64 14.8 82
2006 72,537.12 7.8 62
245 Industry Total 929,033.13 100.0 591
The period of analysis encompasses both the peak of a merger wave and the trough at its beginning.
E.4 Detailed Merger Data
The detailed merger data are provided along with an initial analysis. The columns are described below:
The first three columns provide details of the acquirer, the fourth and fifth
the value and date of the deal, and next three columns the details of the
acquired firm.
There then follows three columns of analysis, namely whether the
acquisition can be traced to one of the Top 48 companies in the analysis
and whether it is a cross-border or cross-sector deal (a ‘1’ signifies that
this is the case).
The next three columns give the values of the deals that are included in
the three cases.
The data in Table E.3 are presented in a table overleaf in landscape format.
246 Table E.3 Detailed Merger Data
Key: Column Title Meaning A Acquirer Name The name of the acquiring company B Acquirer NAIC The NAIC code of the acquiring company C Acquirer Nation The location of the acquiring company D Value ($m) The value of the deal in US$ (million) E Date The date of the transaction F TargetName The name of the acquired company G TargetNAIC The NAIC code of the acquired company H TargetNation The location of the acquired company I Code The abbreviated code given to the acquirer for the subtotal analysis J Ag Value equals 1 if the deal is part of the aggregate total K Xb Value equals 1 if the deal is a cross-border total L Xs Value equals 1 if the deal is a cross-sector total M V(Ag) Value of deal if part of aggregate total N V(Xb) Value of deal if part of cross-border total O V(Xs) Value of deal if part of cross-sector total
247 ABCDEFGHIJKLMNO
Acquirer Name Acquirer NAIC Acquirer Value($m) Date TargetName TargetNAIC Target Code Ag Xb Xs V(Ag) V(Xb) V(Xs) Nation Nation
3M Co Surgical and medical United States 1403 02/08/05 Cuno Inc Fluid power pumps and motors United States instruments and apparatus 3M Co Surgical and medical United States 140 02/03/04 Hornell Ophthalmic goods Sweden instruments and apparatus International AB 3M Co Surgical and medical United States 850 13/12/02 Corning Precision Plastics products, nec United States instruments and apparatus Lens Inc Abbott Laboratories Pharmaceutical preparations United States 320 17/11/04 Experimental & Food preparations, nec United States 1 1 320 320 Applied Science Abbott Laboratories Pharmaceutical preparations United States 1170 06/04/04 TheraSense Inc Surgical and medical instruments United States 1 1 1170 1170 and apparatus Abbott Laboratories Pharmaceutical preparations United States 407 30/01/04 i-Stat Corp Surgical and medical instruments United States 1 1 407 407 and apparatus Abbott Laboratories Pharmaceutical preparations United States 160 27/08/03 ZonePerfect Cereal breakfast foods United States 1 1 160 160 Nutrition Co Abbott Laboratories Pharmaceutical preparations United States 210 30/06/03 Spinal Concepts Orthopedic, prosthetic, and United States 1 1 210 210 Inc surgical supplies Abbott Laboratories Pharmaceutical preparations United States 252 31/05/02 Hokuriku Seiyaku Pharmaceutical preparations Japan 1 1 252 252 Co Ltd Abbott Laboratories Pharmaceutical preparations United States 234 08/05/02 Biocompatibles Int- Pharmaceutical preparations United 1 1 234 234 Cardio Bus Kingdom Abbott Laboratories Pharmaceutical preparations United States 355 05/12/01 Vysis Inc(BP PLC) In vitro and in vivo diagnostic United States 1 355 substances Abbott Laboratories Pharmaceutical preparations United States 6900 02/03/01 Knoll AG(BASF Pharmaceutical preparations Germany 1 1 6900 6900 AG) Abbott Laboratories Pharmaceutical preparations United States 640 19/11/99 Perclose Inc Surgical and medical instruments United States 1 1 640 640 and apparatus Abbott Laboratories Pharmaceutical preparations United States 167 10/07/98 International Murex In vitro and in vivo diagnostic Canada 1 1 167 167 Tech Corp substances Abbott Laboratories Pharmaceutical preparations United States 200 01/05/97 Sanofi Pharmaceutical preparations United States 1 200 Pharmaceuticals- Parente Abbott Laboratories Pharmaceutical preparations United States 802 07/08/96 MediSense Inc In vitro and in vivo diagnostic United States 1 802 substances
248 Abbott Laboratories Pharmaceutical preparations United States 120 14/12/94 Puleva-Nutrition Fluid milk Spain 1 1 1 120 120 120 Division Abbt 14 5 7 11938 7674 3027 Actavis Group hf Pharmaceutical preparations Iceland 810 19/12/05 Alpharma Inc- Pharmaceutical preparations United States Generics Business
Actavis Group hf Pharmaceutical preparations Iceland 600 28/07/05 Amide Pharmaceutical preparations United States Pharmaceutical Inc
Actelion Pharmaceuticals Pharmaceutical preparations Switzerland 191 13/10/03 Axovan AG Pharmaceutical preparations Switzerland Ltd Advanced Medical Inc Pharmaceutical preparations United States 400 26/11/96 IVAC Corp Surgical and medical instruments United States and apparatus Affymetrix Inc Laboratory analytical United States 114 21/10/05 ParAllele Commercial physical and United States instruments BioScience Inc biological research Ajinomoto Co Inc Flavoring extracts and flavoring Japan 183 02/12/02 Shimizu Pharmaceutical preparations Japan syrups, nec Pharmaceutical Co
Akzo Nobel NV Paints, varnishes, lacquers, & Netherlands 711 26/11/99 Hoechst Roussel Pharmaceutical preparations Germany 1 1 1 711 711 711 allied products Vet Akzo Nobel NV Paints, varnishes, lacquers, & Netherlands 3741 07/07/98 Courtaulds PLC Cellulosic manmade fibers United 1 1 3741 3741 allied products Kingdom AkzN 2 2 1 4452 4452 711 Alkermes Inc Biological products, except United States 115 01/02/99 Advanced Pharmaceutical preparations United States diagnostic substances Inhalation Research Allergan Inc Pharmaceutical preparations United States 230 20/11/03 Oculex Surgical and medical instruments United States 1 1 230 230 Pharmaceuticals and apparatus Inc Allergan Inc Pharmaceutical preparations United States 260 16/05/03 Bardeen Sciences Pharmaceutical preparations United States 1 260 Co LLC Allg 2 0 1 490 0 230 Alpharma Inc Pharmaceutical preparations United States 660 12/12/01 FH Faulding & Co- Pharmaceutical preparations United States Oral Pharma Alpharma Inc Pharmaceutical preparations United States 300 02/05/00 Roche Hldg-Animal Pharmaceutical preparations United States Drug Bus Alpharma Inc Pharmaceutical preparations United States 152 18/06/99 Isis Pharma Pharmaceutical preparations Germany GmbH(Schwarz) Alpharma Inc Pharmaceutical preparations United States 198 07/05/98 Arthur H Cox & Co Pharmaceutical preparations United Ltd(Hoechst) Kingdom Altana Chemie AG Chemicals and chemical Germany 769 01/10/05 Eckart GmbH & Co Inorganic pigments Germany preparations, nec KG ALZA Corp Pharmaceutical preparations United States 557 17/03/99 SEQUUS Pharmaceutical preparations United States Pharmaceuticals Inc
249 ALZA Corp Pharmaceutical preparations United States 100 26/08/97 Therapeutic Commercial physical and United States Discovery Corp biological research American Cyanamid Co Chemicals and chemical United States 742 03/06/93 Immunex Corp Biological products, except United States preparations, nec diagnostic substances American Home Products Pharmaceutical preparations United States 449 21/03/97 Solvay Duphar Pharmaceutical preparations Netherlands Corp BV(Solvay SA) American Home Products Pharmaceutical preparations United States 1006 17/12/96 Genetics Institute Pharmaceutical preparations United States Corp Inc American Home Products Pharmaceutical preparations United States 10054 21/12/94 American Chemicals and chemical United States Corp Cyanamid Co preparations, nec American Pacific Corp Chemicals and chemical United States 119 30/11/05 Aerojet Fine Pharmaceutical preparations United States preparations, nec Chemicals LLC American Tropical Plants Medicinal chemicals and United States 105 30/01/98 OPM-USA Inc Radio & TV broadcasting & United States Inc botanical products communications equipment Amersham International Biological products, except United 1345 22/10/97 Nycomed ASA Pharmaceutical preparations Denmark PLC diagnostic substances Kingdom Amersham Life Science Biological products, except United 202 21/09/98 Molecular Laboratory analytical instruments United States diagnostic substances Kingdom Dynamics Inc Amersham Life Science Biological products, except United 373 06/08/97 Pharmacia Biotech Pharmaceutical preparations Sweden diagnostic substances Kingdom AB Amersham PLC Biological products, except United 1000 21/03/02 Amersham Biological products, except Sweden diagnostic substances Kingdom Biosciences AB diagnostic substances Amgen Inc Biological products, except United States 1285 13/08/04 Tularik Inc Pharmaceutical preparations United States 1 1285 diagnostic substances Amgen Inc Biological products, except United States 16685 16/07/02 Immunex Corp Biological products, except United States 1 16685 diagnostic substances diagnostic substances Amgen Inc Biological products, except United States 138 31/05/02 Roche-Filgrastm & Pharmaceutical preparations Switzerland 1 1 138 138 diagnostic substances Pegrilgrastm Amgen Inc Biological products, except United States 169 14/12/00 Kinetix Biological products, except United States 1 169 diagnostic substances Pharmaceuticals diagnostic substances Inc Amgn 4 1 0 18276 138 0 Arch Chemicals Inc Chemicals and chemical United States 219 05/04/04 Avecia Inc Chemicals and chemical United States preparations, nec preparations, nec Arch Chemicals Inc Chemicals and chemical United States 184 22/08/00 Hickson Plastics materials and synthetic United preparations, nec International PLC resins Kingdom Arris Pharmaceuticals Corp Pharmaceutical preparations United States 170 09/01/98 Sequana In vitro and in vivo diagnostic United States Therapeutics substances Asahi Breweries Ltd Malt beverages Japan 151 02/09/02 Kyowa Hakko Beer and ale Japan Kogyo-Alcohol Sale
Astra AB Pharmaceutical preparations Sweden 6090 01/07/98 Astra Merck Drugs, drug proprietaries, and United States 1 1 6090 6090 Inc(Merck & Co) druggists' sundries Astra AB Pharmaceutical preparations Sweden 320 16/05/95 Fisons PLC- Commercial physical and United 1 1 320 320 Pharmaceutical biological research Kingdom
250 ZENECA Group PLC Pharmaceutical preparations United 31774 06/04/99 Astra AB Pharmaceutical preparations Sweden 1 1 31774 31774 Kingdom ZENECA Group PLC Pharmaceutical preparations United 193 04/09/98 Orica Ltd-Pharm Pharmaceutical preparations Australia 1 1 193 193 Kingdom Business ZENECA Group PLC Pharmaceutical preparations United 410 04/02/98 Ishihara Sangyo Pesticides and agricultural United States 1 1 1 410 410 410 Kingdom Kaisha Ltd-US chemicals, nec ZENECA Group PLC Pharmaceutical preparations United 234 14/04/97 Salick Health Care Kidney dialysis centers United States 1 1 1 234 234 234 Kingdom Inc AstrZ 6 6 2 39021 39021 644 Axcan Pharma Inc Pharmaceutical preparations Canada 145 18/11/03 Aventis SA- Pharmaceutical preparations United States Carafete,4 Others Axcan Pharma Inc Pharmaceutical preparations Canada 108 30/09/99 Scandipharm Inc Drugs, drug proprietaries, and United States druggists' sundries Barr Laboratories Inc Pharmaceutical preparations United States 638 24/10/01 Duramed Pharmaceutical preparations United States Pharmaceuticals Inc Bausch & Lomb Inc Ophthalmic goods United States 200 26/09/05 Sino Concept Investors, nec Hong Kong 1 1 1 200 200 200 Technology Ltd Bausch & Lomb Inc Ophthalmic goods United States 227 08/08/00 Chauvin Pharmaceutical preparations France 1 1 227 227 Bausch & Lomb Inc Ophthalmic goods United States 380 05/01/98 Storz Instrument Surgical and medical instruments United States 1 1 380 380 Co and apparatus Bausch & Lomb Inc Ophthalmic goods United States 300 29/12/97 Chiron Surgical and medical instruments United States 1 1 300 300 Vision(Chiron and apparatus Corp) Bausch & Lomb Inc Ophthalmic goods United States 129 02/08/93 Dahlberg Inc Orthopedic, prosthetic, and United States 1 1 129 129 surgical supplies BausL 5 2 4 1237 427 1009 Baxter Healthcare Corp Pharmaceutical preparations United States 305 20/12/02 Wyeth-Certain ESI Pharmaceutical preparations United States 1 305 Lederle Asts Baxter Healthcare Corp Pharmaceutical preparations United States 219 20/08/01 Cook Pharmaceutical preparations United States 1 219 Pharmaceutical Solutions Baxter International Inc Biological products, except United States 148 05/05/02 Fusion Medical Surgical and medical instruments United States 1 1 148 148 diagnostic substances Technologies and apparatus Baxter International Inc Biological products, except United States 396 26/06/00 North American Biological products, except United States 1 396 diagnostic substances Vaccine Inc diagnostic substances Baxter International Inc Biological products, except United States 182 07/03/00 Althin Medical AB Surgical and medical instruments Sweden 1 1 1 182 182 182 diagnostic substances and apparatus Baxter International Inc Biological products, except United States 189 04/05/98 Somatogen Inc Biological products, except United States 1 189 diagnostic substances diagnostic substances Baxter International Inc Biological products, except United States 104 03/04/98 Ohmeda- Medicinal chemicals and botanical United States 1 104 diagnostic substances Pharmaceutical products Prod Div Baxter International Inc Biological products, except United States 235 31/03/98 Bieffe Medital SpA- Electromedical and Switzerland 1 1 1 235 235 235 diagnostic substances Dialysis electrotherapeutic apparatus
251 Baxter International Inc Biological products, except United States 236 17/03/97 Research Medical Surgical and medical instruments United States 1 1 236 236 diagnostic substances Inc and apparatus Baxter International Inc Biological products, except United States 213 17/02/97 Immuno Pharmaceutical preparations Switzerland 1 1 213 213 diagnostic substances International AG Baxter International Inc Biological products, except United States 206 30/01/97 Immuno Pharmaceutical preparations Switzerland 1 1 206 206 diagnostic substances International AG Baxter International Inc Biological products, except United States 219 19/12/96 Immuno Pharmaceutical preparations Switzerland 1 1 219 219 diagnostic substances International AG Baxt 12 5 4 2652 1055 801 Bayer AG Medicinal chemicals and Germany 2961 03/01/05 Roche Holding AG- Pharmaceutical preparations Switzerland 1 1 2961 2961 botanical products Over-The Bayer AG Medicinal chemicals and Germany 6646 03/06/02 Aventis Pesticides and agricultural France 1 1 1 6646 6646 6646 botanical products CropScience Hldg chemicals, nec SA Bayer AG Medicinal chemicals and Germany 106 01/02/01 Syngenta AG- Pesticides and agricultural Switzerland 1 1 1 106 106 106 botanical products Mikado Herbicide chemicals, nec Bayer AG Medicinal chemicals and Germany 327 24/10/00 Sybron Chemicals Chemicals and chemical United States 1 1 1 327 327 327 botanical products Inc preparations, nec Bayer AG Medicinal chemicals and Germany 2450 31/03/00 Lyondell Chemical- Petroleum refining United States 1 1 1 2450 2450 2450 botanical products Polyils Bus Bayer AG Medicinal chemicals and Germany 1100 30/11/98 Chiron Diagnostics Pharmaceutical preparations United States 1 1 1100 1100 botanical products Corp Bayer AG Medicinal chemicals and Germany 580 02/01/96 Monsanto Co- Plastics products, nec United States 1 1 1 580 580 580 botanical products Styrenics Plastics Bayer(India)Ltd Pharmaceutical preparations India 360 16/05/03 Bayer CropScience Pesticides and agricultural India 1 1 1 360 360 360 India Ltd chemicals, nec
Bayr 8 8 6 14530 14530 10469 Becton Dickinson & Co Surgical and medical United States 195 26/08/99 Clontech Biological products, except United States instruments and apparatus Laboratories Inc diagnostic substances Becton Dickinson & Co Surgical and medical United States 452 03/04/98 Ohmeda-Medical Surgical and medical instruments United States instruments and apparatus Devices Div and apparatus Berna Biotech AG Biological products, except Switzerland 234 05/08/02 Rhein Biotech NV Biological products, except Netherlands diagnostic substances diagnostic substances Berna Biotech AG Biological products, except Switzerland 110 04/03/00 Green Cross Biological products, except South Korea diagnostic substances Vaccine Corp diagnostic substances BioMarin Pharmaceutical Pharmaceutical preparations United States 190 18/05/04 Ascent Pediatrics Pharmaceutical preparations United States Inc Inc BioMarin Pharmaceutical Pharmaceutical preparations United States 141 22/08/02 Glyko Biomedical In vitro and in vivo diagnostic United States Inc Ltd substances bioMerieux Pierre Fabre Surgical and medical France 285 03/07/01 Organon Tek-In In vitro and in vivo diagnostic Netherlands instruments and apparatus Vitro Diagn Bus substances Bio-Rad Laboratories Inc Laboratory analytical United States 210 04/10/99 Pasteur Sanofi Medicinal chemicals and botanical France instruments Diagnostics products
252 Biovail Corp Pharmaceutical preparations Canada 130 02/06/03 Wyeth-Ativan & Pharmaceutical preparations United States Isordil Rights Biovail Corp Pharmaceutical preparations Canada 190 11/12/02 Pharma PASS LLC Pharmaceutical preparations United States
Biovail Corp Pharmaceutical preparations Canada 410 29/12/01 Aventis-Product Pharmaceutical preparations United States Line Biovail Corp Pharmaceutical preparations Canada 213 06/10/00 DJ Pharma Pharmaceutical preparations United States Biovail Corp International Pharmaceutical preparations Canada 166 12/11/99 Fuisz Technologies Pharmaceutical preparations United States Ltd BOC Group PLC Industrial gases United 109 12/07/93 Huels AG- Industrial gases Germany Kingdom Hydrogen Business Boots Co PLC Pharmaceutical preparations United 340 07/12/00 Procter & Gamble- Pharmaceutical preparations United States Kingdom Clearasil Boots Co PLC Pharmaceutical preparations United 278 01/10/97 Hermal Kurt Pharmaceutical preparations Germany Kingdom Herrman(Merck E) Boots Healthcare Pharmaceutical preparations United 179 26/09/96 Lutsia(Roussel- Perfumes, cosmetics, and other France International Kingdom Uclaf/Hoechst) toilet preparations Boots Healthcare Pharmaceutical preparations United 179 20/09/96 Laboratoires Perfumes, cosmetics, and other France International Kingdom Lutsia(Roussel) toilet preparations Bracco SpA Pharmaceutical preparations Italy 881 22/03/00 Merck,Bracco- X-Ray apparatus & tubes & other Italy Contrast Imaging irradiation equip. Bradley Pharmaceuticals Pharmaceutical preparations United States 183 10/08/04 Bioglan Pharma Pharmaceutical preparations United States Inc Inc Bristol-Myers Squibb Co Pharmaceutical preparations United States 7800 02/10/01 DuPont Pharmaceutical preparations United States 1 7800 Pharmaceuticals Co Bristol-Myers Squibb Co Pharmaceutical preparations United States 150 11/03/96 Argentia SA Pharmaceutical preparations Argentina 1 1 150 150 Bristol-Myers Squibb Co Pharmaceutical preparations United States 262 04/01/95 Calgon Vestal Medicinal chemicals and botanical United States 1 262 Laboratories products BrMS 3 1 0 8212 150 0 Cambrex Corp Pharmaceutical preparations United States 145 04/06/01 Bio Science Medicinal chemicals and botanical United States Contract Prodn Cor products
Cambrex Corp Industrial organic chemicals, United States 132 03/10/97 BioWhittaker Inc In vitro and in vivo diagnostic United States nec substances Cambrex Corp Industrial organic chemicals, United States 130 12/10/94 Akzo Nobel-Nobel Medicinal chemicals and botanical Netherlands nec Pharma products Cargill Inc Soybean oil mills United States 284 12/04/05 Seara Alimentos Sausages and other prepared Brazil SA meat products Cargill Inc Soybean oil mills United States 1068 10/05/02 Cerestar Wet corn milling France Cargill Inc Soybean oil mills United States 429 04/04/02 Cerestar Wet corn milling France Cargill Inc Soybean oil mills United States 440 30/04/01 Agribrands Prepared animal feeds, except for United States International Inc dogs and cats
253 Cargill Inc Soybean oil mills United States 140 02/12/98 Grandes Molinos Flour and other grain mill products Venezuela de Venezuela Celera Genomics Corp Commercial physical and United States 140 16/11/01 AXYS Pharmaceutical preparations United States biological research Pharmaceuticals Inc Celgene Corp Pharmaceutical preparations United States 110 21/10/04 Penn T Pharmaceutical preparations United Kingdom Celgene Corp Pharmaceutical preparations United States 198 01/09/00 Signal Commercial physical and United States Pharmaceuticals biological research Inc Cell Pathways Holdings Inc Pharmaceutical preparations United States 151 03/11/98 Tseng Computer peripheral equipment, United States Laboratories Inc nec Cell Therapeutics Inc Pharmaceutical preparations United States 137 02/01/04 Novuspharma SpA Pharmaceutical preparations Italy Centocor Inc In vitro and in vivo diagnostic United States 335 24/03/98 Roche Healthcare- Drugs, drug proprietaries, and United States substances Centocor Mktg druggists' sundries Cephalon Inc Pharmaceutical preparations United States 360 22/12/05 Zeneus Holdings Pharmaceutical preparations United 1 1 360 360 Ltd Kingdom Cephalon Inc Pharmaceutical preparations United States 170 19/07/05 CTI Technologies Pharmaceutical preparations United States 1 170 Inc-Trisenox Cephalon Inc Pharmaceutical preparations United States 150 14/06/05 Salmedix Inc Commercial physical and United States 1 150 biological research Cephalon Inc Pharmaceutical preparations United States 430 12/08/04 CIMA Labs Inc Pharmaceutical preparations United States 1 430 Cephalon Inc Pharmaceutical preparations United States 450 28/12/01 Laboratoire L Pharmaceutical preparations France 1 1 450 450 Lafon Cephalon Inc Pharmaceutical preparations United States 438 11/10/00 Anesta Corp Biological products, except United States 1 438 diagnostic substances Ceph 6 2 0 1998 810 0 Chattem Inc Pharmaceutical preparations United States 165 24/03/98 Bristol-Myers-Ban Perfumes, cosmetics, and other United States Anti-Perspir toilet preparations Chiron Corp Biological products, except United States 789 08/07/03 PowderJect Biological products, except United diagnostic substances Pharmaceuticals diagnostic substances Kingdom PLC Chiron Corp Biological products, except United States 699 22/09/00 PathoGenesis Biological products, except United States diagnostic substances Corp diagnostic substances Chiron Corp Biological products, except United States 125 01/04/98 Behringwerke AG- Biological products, except Germany diagnostic substances Human Vaccine diagnostic substances Chiron Corp Biological products, except United States 110 31/03/98 Chiron Behring Biological products, except Germany diagnostic substances GmbH & Co diagnostic substances Chiron Corp Biological products, except United States 112 02/10/95 Viagene Inc Commercial physical and United States diagnostic substances biological research Chiron Corp Biological products, except United States 616 05/01/95 Ciba-Corning Pharmaceutical preparations United States diagnostic substances Diag,Biocine Chiroscience Group PLC Pharmaceutical preparations United 112 19/12/96 Darwin Molecular Commercial physical and United States Kingdom Corp biological research
254 Christian Hansen Holding Food preparations, nec Denmark 103 09/12/98 Ingredients Food preparations, nec United States A/S Technology Corp Chugai Pharmaceutical Co Pharmaceutical preparations Japan 2590 01/10/02 Nippon Roche KK Pharmaceutical preparations Japan 1 2590 Ltd Chug 1 0 0 2590 0 0 Ciba Specialty Chemicals Chemicals and chemical Switzerland 584 03/06/04 Raisio Chemicals Industrial inorganic chemicals, nec Finland preparations, nec Oy Ciba Specialty Chemicals Chemicals and chemical Switzerland 2501 12/03/98 Allied Colloids Industrial organic chemicals, nec United preparations, nec Group PLC Kingdom Ciba-Geigy AG Pharmaceutical preparations Switzerland 357 22/12/94 Rhone-Poulenc Pharmaceutical preparations United States Rorer-US and Can Ciba-Geigy AG Pharmaceutical preparations Switzerland 140 07/01/93 Fisons PLC-North Pharmaceutical preparations United States American Connetics Corp Pharmaceutical preparations United States 123 04/03/04 Hoffman-Soriatane Pharmaceutical preparations United States Rights Cooper Cos Inc Ophthalmic goods United States 1130 06/01/05 Ocular Sciences Ophthalmic goods United States Inc Cordis Corp Surgical and medical United States 400 16/10/97 Biosense Inc Electromedical and Israel instruments and apparatus electrotherapeutic apparatus Corgentech Inc Biological products, except United States 130 15/12/05 AlgoRx Pharmaceutical preparations United States diagnostic substances Pharmaceuticals Inc Corixa Corp Biological products, except United States 819 22/12/00 Coulter Biological products, except United States diagnostic substances Pharmaceuticals diagnostic substances Inc Creative BioMolecules Inc Biological products, except United States 104 01/08/00 Ontogeny Inc Health and allied services, nec United States diagnostic substances CSL Ltd Pharmaceutical preparations Australia 925 31/03/04 Aventis Behring Biological products, except United States 1 1 925 925 LLC diagnostic substances CSL Ltd Pharmaceutical preparations Australia 152 08/09/01 Nabi Inc-Plasma Biological products, except United States 1 1 152 152 Collection diagnostic substances CSL Ltd Pharmaceutical preparations Australia 592 30/08/00 ZLB Central Biological products, except Switzerland 1 1 592 592 Laboratory Blood diagnostic substances CSL 3 3 0 1669 1669 0 Dade International Inc In vitro and in vivo diagnostic United States 525 08/05/96 EI du Pont de Inorganic pigments United States substances Nemmours-In Sankyo Co Ltd Pharmaceutical preparations Japan 6290 28/09/05 Daiichi Pharmaceutical preparations Japan 1 6290 Pharmaceutical Co Ltd Daic 1 0 0 6290 0 0 Dainippon Pharm Co Ltd Pharmaceutical preparations Japan 2224 01/10/05 Sumitomo Pharmaceutical preparations Japan 1 2224 Pharmaceuticals Co Dain 1 0 0 2224 0 0
255 Diosynth BV Pharmaceutical preparations Netherlands 190 15/06/01 Covance Pharmaceutical preparations United States Biotechnology Services Dow Italia Spa(Dow Medicinal chemicals and Italy 300 04/01/96 INCA Intl(Enichem Custom compounding of Italy Chemicals) botanical products SpA/ENI/IT) purchased plastics resins Dr Reddy's Laboratories Pharmaceutical preparations India 211 01/04/00 Cheminor Drugs Pharmaceutical preparations India Ltd Ltd DSM NV Chemicals and chemical Netherlands 686 02/02/05 NeoResins Plastics materials and synthetic Netherlands preparations, nec resins DSM NV Chemicals and chemical Netherlands 1915 30/09/03 Roche Holding AG- Medicinal chemicals and botanical Switzerland preparations, nec Vitamins products DSM NV Chemicals and chemical Netherlands 800 14/12/00 Catalytica Industrial inorganic chemicals, nec United States preparations, nec Pharmaceuticals Inc DSM NV Chemicals and chemical Netherlands 1729 11/05/98 Koninklijke Gist- Industrial organic chemicals, nec Netherlands preparations, nec Brocades NV Duramed Pharmaceuticals Pharmaceutical preparations United States 282 10/11/05 FEI Womens Medical, dental, and hospital United States Inc Health LLC equipment & supplies Duramed Pharmaceuticals Pharmaceutical preparations United States 142 08/09/05 Organon Pharm Pharmaceutical preparations United States Inc USA Inc-Mircette Eisai Co Ltd Pharmaceutical preparations Japan 265 27/04/04 Elan Corp- Pharmaceutical preparations United States 1 1 265 265 Zonegan Rights Eisa 1 1 0 265 265 0 Elan Corp PLC Biological products, except Ireland-Rep 1860 10/11/00 Dura Pharmaceutical preparations United States diagnostic substances Pharmaceuticals Inc Elan Corp PLC Biological products, except Ireland-Rep 601 15/05/00 Liposome Co Inc Pharmaceutical preparations United States diagnostic substances Elan Corp PLC Biological products, except Ireland-Rep 183 31/12/99 Axogen Ltd Pharmaceutical preparations Bermuda diagnostic substances Elan Corp PLC Biological products, except Ireland-Rep 150 01/10/98 NanoSystems Pharmaceutical preparations United States diagnostic substances LLC(Eastman Kodak) Elan Corp PLC Biological products, except Ireland-Rep 773 14/08/98 Neurex Corp Biological products, except United States diagnostic substances diagnostic substances Elan Corp PLC Biological products, except Ireland-Rep 150 01/06/98 Carnrick Drugs, drug proprietaries, and United States diagnostic substances Laboratories Inc druggists' sundries Elan Corp PLC Biological products, except Ireland-Rep 398 02/03/98 Sano Corp Pharmaceutical preparations United States diagnostic substances Elan Corp PLC Biological products, except Ireland-Rep 141 30/10/96 Advanced Pharmaceutical preparations Bermuda diagnostic substances Therapeutic Systems Elan Corp PLC Biological products, except Ireland-Rep 576 01/07/96 Athena Pharmaceutical preparations United States diagnostic substances Neurosciences Inc Elders Australia Ltd Farm management services Australia 207 28/10/93 Elders Ltd Farm management services Australia
256 Eli Lilly & Co Pharmaceutical preparations United States 381 12/02/04 Applied Molecular Biological products, except United States 1 381 Evolution diagnostic substances Eli Lilly & Co Pharmaceutical preparations United States 4000 21/11/94 PCS Health Data processing services United States 1 4000 Systems EliL 2 0 0 4381 0 0 Enaleni Pharmaceutical preparations South Africa 186 31/10/05 Cipla Pharmaceutical preparations South Africa Pharmaceuticals(Pty) Medpro(Pty)Ltd Enzon Inc Biological products, except United States 360 22/11/02 Elan Corp Plc- Pharmaceutical preparations United States diagnostic substances Abelcet Rights & Epitope Inc In vitro and in vivo diagnostic United States 255 28/09/00 STC Technologies Pharmaceutical preparations United States substances Inc Ercros SA Chemicals and chemical Spain 218 02/06/05 Uralita SA- Chemicals and chemical Spain preparations, nec Chemical Divisions preparations, nec
Evotech BioSystems AG Pharmaceutical preparations Germany 459 29/09/00 Oxford Asymmetry Pharmaceutical preparations United International Kingdom
Exelixis Inc Biological products, except United States 104 08/01/02 Genomica Corp Computer programming services United States diagnostic substances Fidia Farmaceutici SpA Pharmaceutical preparations Italy 186 29/05/03 Antibioticos SA Pharmaceutical preparations Spain Fisher Scientific Intl Inc Surgical and medical United States 150 03/08/05 Lancaster Commercial physical and United States instruments and apparatus Laboratories Inc biological research Fisher Scientific Intl Inc Surgical and medical United States 3669 02/08/04 Apogent Laboratory apparatus and United States instruments and apparatus Technologies Inc furniture Fisher Scientific Intl Inc Surgical and medical United States 330 01/03/04 Oxoid Holdings Ltd In vitro and in vivo diagnostic United instruments and apparatus substances Kingdom Fisher Scientific Intl Inc Surgical and medical United States 786 03/09/03 Perbio Science AB Surgical and medical instruments Sweden instruments and apparatus and apparatus Fisher Scientific Intl Inc Surgical and medical United States 205 05/11/01 Cole-Parmer Chemicals and allied products, United States instruments and apparatus Instrument Co nec Fisher Scientific Intl Inc Surgical and medical United States 138 15/02/01 Covance Inc- Packing and crating United States instruments and apparatus Pharmaceutical Fisher Scientific Intl Inc Surgical and medical United States 310 17/10/95 Fisons Scientific Medical, dental, and hospital United instruments and apparatus Equip,Curtin equipment & supplies Kingdom Fresenius AG Pharmaceutical preparations Germany 472 11/12/98 Pharmacia & Pharmaceutical preparations Sweden Upjohn-Nutrition Fresenius AG Pharmaceutical preparations Germany 4236 30/09/96 National Medical Medical, dental, and hospital United States Care Inc equipment & supplies Fujirebio Inc Pharmaceutical preparations Japan 168 11/11/04 SRL Inc Commercial physical and Japan biological research Fukujin Co Ltd Pharmaceutical preparations Japan 205 29/09/03 Azwell Inc Pharmaceutical preparations Japan Galen Holdings PLC Pharmaceutical preparations United 484 27/03/03 Pfizer Inc- Pharmaceutical preparations United States Kingdom Estrostep,Loestrin
257 Galen Holdings PLC Pharmaceutical preparations United 295 23/01/03 Eli Lilly- Pharmaceutical preparations United States Kingdom Sales,Marketing Rts Galen Holdings PLC Pharmaceutical preparations United 551 28/09/00 Warner Chilcott Pharmaceutical preparations Ireland-Rep Kingdom PLC Genentech Inc Biological products, except United States 408 24/06/05 Biogen Idec- Biological products, except United States 1 408 diagnostic substances Biologics Mnfr Fac diagnostic substances Gene 1 0 0 408 0 0 Genome Therapeutics Biological products, except United States 104 06/02/04 Genesoft Inc Biological products, except United States Corp diagnostic substances diagnostic substances Gensia Sicor Inc Pharmaceutical preparations United States 140 28/02/97 Rakepoll Holding Offices of holding companies, nec Netherlands 1 140 BV(Rakepoll) Genzyme Biosurgery Biological products, except United States 427 18/12/00 Biomatrix Inc Medicinal chemicals and botanical United States 1 427 diagnostic substances products Genzyme Corp Biological products, except United States 595 01/07/05 Bone Care Intl Inc Pharmaceutical preparations United States 1 595 diagnostic substances Genzyme Corp Biological products, except United States 415 06/01/05 Wyeth- Pharmaceutical preparations United States 1 415 diagnostic substances Sales,Marketing Rights Genzyme Corp Biological products, except United States 949 21/12/04 ILEX Oncology Inc Pharmaceutical preparations United States 1 949 diagnostic substances Genzyme Corp Biological products, except United States 215 03/05/04 Impath Physician Commercial nonphysical research United States 1 215 diagnostic substances Services Genzyme Corp Biological products, except United States 535 15/09/03 SangStat Medical Pharmaceutical preparations United States 1 535 diagnostic substances Corp Genzyme Corp Biological products, except United States 225 27/09/01 Novazyme Pharmaceutical preparations United States 1 225 diagnostic substances Pharmaceuticals Inc Genzyme Corp Biological products, except United States 993 14/12/00 GelTex Biological products, except United States 1 993 diagnostic substances Pharmaceuticals diagnostic substances Inc Genzyme Corp Biological products, except United States 107 29/10/96 Neozyme II Corp Pharmaceutical preparations British Virgin 1 1 107 107 diagnostic substances Genzyme Corp Biological products, except United States 250 02/07/96 Deknatel Snowden Surgical and medical instruments United States 1 1 250 250 diagnostic substances Pencer and apparatus Genz 10 1 1 4710 107 250 Gilead Sciences Inc Biological products, except United States 123 15/09/03 Equity Office- Colleges, universities, and United States 1 123 diagnostic substances Foster City professional schools Gilead Sciences Inc Biological products, except United States 407 23/01/03 Triangle Pharmaceutical preparations United States 1 407 diagnostic substances Pharmaceuticals Inc Gilead Sciences Inc Biological products, except United States 866 29/07/99 NeXstar Pharmaceutical preparations United States 1 866 diagnostic substances Pharmaceuticals Inc Gild 3 0 0 1396 0 0
258 Glaxo Holdings PLC Pharmaceutical preparations United 605 21/11/96 Nippon Glaxo Pharmaceutical preparations Japan 1 1 605 605 Kingdom Glaxo Holdings PLC Pharmaceutical preparations United 13408 16/03/95 Wellcome PLC Pharmaceutical preparations United 1 13408 Kingdom Kingdom Glaxo Wellcome PLC Pharmaceutical preparations United 78775 27/12/00 SmithKline Medicinal chemicals and botanical United 1 78775 Kingdom Beecham PLC products Kingdom Glaxo Wellcome PLC Pharmaceutical preparations United 106 08/01/99 Amoun Pharmaceutical preparations Egypt 1 1 106 106 Kingdom Pharmaceuticals {APIC} Glaxo Wellcome PLC Pharmaceutical preparations United 220 28/01/98 Polfa Pharmaceutical preparations Poland 1 1 220 220 Kingdom Poznan(Poland) GlaxoSmithKline PLC Pharmaceutical preparations United 1388 08/12/05 ID Biomedical Corp Biological products, except Canada 1 1 1388 1388 Kingdom diagnostic substances GlaxoSmithKline PLC Pharmaceutical preparations United 349 12/07/05 Corixa Corp Biological products, except United States 1 1 349 349 Kingdom diagnostic substances GlaxoSmithKline PLC Pharmaceutical preparations United 547 03/09/04 Sanofi-Synthelabo- Pharmaceutical preparations France 1 1 547 547 Kingdom Drugs SmithKline Beecham Corp Pharmaceutical preparations United States 2300 27/05/94 Diversified Offices and clinics of doctors of United States 1 1 2300 2300 Pharmaceutical medicine SmithKline Beecham PLC Medicinal chemicals and United 1453 16/01/01 Block Drug Co Dental equipment and supplies United States 1 1 1 1453 1453 1453 botanical products Kingdom SmithKline Beecham PLC Medicinal chemicals and United 141 05/01/96 Abtei Pharma- Drugs, drug proprietaries, and Germany 1 1 141 141 botanical products Kingdom Vertriebs GmbH druggists' sundries SmithKline Beecham PLC Medicinal chemicals and United 2925 02/11/94 Sterling Winthrop Pharmaceutical preparations United States 1 1 2925 2925 botanical products Kingdom Inc GSK 12 10 1 102218 10035 1453 Global Pharm Dvlp Inc Biological products, except United States 125 30/09/05 Quintiles-Business Biological products, except United States diagnostic substances Units(3) diagnostic substances Global Pharmaceutical Pharmaceutical preparations United States 139 15/12/99 Impax Pharmaceutical preparations United States Corp Pharmaceuticals Inc Grasim Industries Ltd Pulp mills India 275 06/07/04 Larsen & Toubro Cement, hydraulic India Ltd-Cement Guidant Corp Surgical and medical United States 291 15/11/99 CardioThoracic Electromedical and United States instruments and apparatus Systems Inc electrotherapeutic apparatus Guidant Corp Surgical and medical United States 810 01/02/99 SulzerMedica- Electromedical and United States instruments and apparatus Electrophysiology electrotherapeutic apparatus Guidant Corp Surgical and medical United States 121 31/12/98 InControl Inc Orthopedic, prosthetic, and United States instruments and apparatus surgical supplies Guidant Corp Surgical and medical United States 190 19/12/97 EndoVascular Surgical and medical instruments United States instruments and apparatus Technologies Inc and apparatus H Lundbeck A/S Pharmaceutical preparations Denmark 135 06/03/03 Synaptic Pharmaceutical preparations United States 1 1 135 135 Pharmaceutical Corp H Lundbeck A/S Pharmaceutical preparations Denmark 101 02/02/01 Lundbeck GmbH Pharmaceutical preparations Germany 1 1 101 101
259 Hlun 2 0 2 236 0 236 Hafslund Nycomed AS Pharmaceutical preparations Norway 450 03/10/94 Sterling Winthrop- Electromedical and United States Med Image electrotherapeutic apparatus Herba-Apotheker AG Medicinal chemicals and Austria 125 12/09/97 Chemosan-Union Medicinal chemicals and botanical Austria botanical products AG products Hisamitsu Pharmaceutical Pharmaceutical preparations Japan 136 01/04/05 Biomedics Pharmaceutical preparations Japan Human Genome Sciences In vitro and in vivo diagnostic United States 120 08/09/00 Principia Biological products, except United States Inc substances Pharmaceutical diagnostic substances Corp ICN Pharmaceuticals Inc Pharmaceutical preparations United States 187 22/08/03 Ribapharm Inc Biological products, except United States diagnostic substances ID Biomedical Corp Biological products, except Canada 116 09/09/04 Shire Biologics Biological products, except Canada diagnostic substances diagnostic substances IDEC Pharmaceuticals Biological products, except United States 6059 12/11/03 Biogen Inc Biological products, except United States Corp diagnostic substances diagnostic substances Immunex Corp Biological products, except United States 468 01/01/02 Greenwich Pharmaceutical preparations United States diagnostic substances Holdings Inc Inhale Therapeutic Pharmaceutical preparations United States 191 29/06/01 Shearwater Corp Pharmaceutical preparations United States Systems Inc Inhale Therapeutic Pharmaceutical preparations United States 200 09/01/01 Bradford Particle Commercial physical and United Systems Inc Design PLC biological research Kingdom Intercare Group PLC Pharmaceutical preparations United 122 26/10/00 Macarthy Group Pharmaceutical preparations United Kingdom Ltd(Cinven) Kingdom Inverness Med Innovations In vitro and in vivo diagnostic United States 149 20/12/01 Unipath Surgical and medical instruments United Inc substances Ltd(Unilever PLC) and apparatus Kingdom Invitrogen Corp Biological products, except United States 131 06/10/05 BioSource In vitro and in vivo diagnostic United States diagnostic substances International Inc substances Invitrogen Corp Biological products, except United States 388 01/04/05 Dynal Biotech ASA Biological products, except Norway diagnostic substances diagnostic substances Invitrogen Corp Biological products, except United States 486 10/02/04 BioReliance Corp Commercial physical and United States diagnostic substances biological research Invitrogen Corp Biological products, except United States 325 22/08/03 Molecular Probes Chemicals and allied products, United States diagnostic substances Inc nec Invitrogen Corp Biological products, except United States 402 14/09/00 Life Technologies Biological products, except United States diagnostic substances Inc(Dexter) diagnostic substances Invitrogen Corp Biological products, except United States 1660 14/09/00 Dexter Corp Adhesives and sealants United States diagnostic substances Invitrogen Corp Biological products, except United States 127 02/02/00 Research Genetics Commercial physical and United States diagnostic substances Inc biological research Ion Beam Applications SA Electromedical and Belgium 225 22/07/99 Sterigenics Business services, nec United States electrotherapeutic apparatus International Inc IVAX Corp Pharmaceutical preparations United States 272 11/05/05 Phoenix Scientific Pharmaceutical preparations United States Inc IVAX Corp Pharmaceutical preparations United States 453 29/06/01 Laboratorio Chile Pharmaceutical preparations Chile SA
260 IVAX Corp Pharmaceutical preparations United States 605 30/12/94 Zenith Pharmaceutical preparations United States Laboratories Inc IVAX Corp Pharmaceutical preparations United States 585 28/03/94 McGaw Inc Pharmaceutical preparations United States Jazz Pharmaceuticals Inc Pharmaceutical preparations United States 131 27/06/05 Orphan Medical Pharmaceutical preparations United States Inc Johnson & Johnson Pharmaceutical preparations United States 387 03/06/05 Closure Medical Surgical and medical instruments United States 1 1 387 387 Corp and apparatus Johnson & Johnson Pharmaceutical preparations United States 230 04/04/05 TransForm Pharmaceutical preparations United States 1 230 Pharmaceuticals Inc Johnson & Johnson Pharmaceutical preparations United States 2449 29/04/03 Scios Inc Pharmaceutical preparations United States 1 2449 Johnson & Johnson Pharmaceutical preparations United States 320 18/04/02 Tibotec-Virco NV Medical laboratories Belgium 1 1 320 320 Johnson & Johnson Pharmaceutical preparations United States 1300 21/11/01 Inverness Medical- Electromedical and United States 1 1 1300 1300 Diabetes electrotherapeutic apparatus Johnson & Johnson Pharmaceutical preparations United States 10213 22/06/01 ALZA Corp Pharmaceutical preparations United States 1 10213 Johnson & Johnson Pharmaceutical preparations United States 4861 06/10/99 Centocor Inc In vitro and in vivo diagnostic United States 1 4861 substances Johnson & Johnson Pharmaceutical preparations United States 3360 05/11/98 Depuy Inc(Corange Orthopedic, prosthetic, and United States 1 3360 Ltd) surgical supplies Johnson & Johnson Pharmaceutical preparations United States 296 31/07/97 Biopsys Medical Surgical and medical instruments United States 1 1 296 296 Inc and apparatus Johnson & Johnson Pharmaceutical preparations United States 118 24/03/97 Innotech Inc Electromedical and United States 1 1 118 118 electrotherapeutic apparatus Johnson & Johnson Pharmaceutical preparations United States 1789 23/02/96 Cordis Corp X-Ray apparatus & tubes & other United States 1 1 1789 1789 irradiation equip. Johnson & Johnson Pharmaceutical preparations United States 124 05/04/95 Mitek Surgical Surgical and medical instruments United States 1 1 124 124 Products and apparatus Johnson & Johnson Pharmaceutical preparations United States 1008 30/11/94 Eastman Kodak- In vitro and in vivo diagnostic United States 1 1008 Clinical substances Johnson & Johnson Pharmaceutical preparations United States 900 03/10/94 Neutrogena Corp Soap & other detergents, except United States 1 1 900 900 specialty cleaners Johnson & Johnson Pharmaceutical preparations United States 169 09/12/93 Roc(LVMH-Moet Perfumes, cosmetics, and other France 1 1 1 169 169 169 Hennessy L Vuit) toilet preparations John 15 2 8 27524 489 5083 Johnson Matthey PLC Chemicals and chemical United 404 01/11/02 ICI Synetix Industrial inorganic chemicals, nec United preparations, nec Kingdom Kingdom Johnson Matthey PLC Chemicals and chemical United 206 09/07/01 Meconic PLC Drugs, drug proprietaries, and United preparations, nec Kingdom druggists' sundries Kingdom Johnson Matthey PLC Chemicals and chemical United 216 06/02/98 Cookson Matthey Pottery products, nec United preparations, nec Kingdom Ceramics and Kingdom Johnson Matthey PLC Chemicals and chemical United 164 06/10/95 Advance Circuits Printed circuit boards United States preparations, nec Kingdom Inc Kalbe Farma PT Pharmaceutical preparations Indonesia 473 20/12/05 Enseval Drugs, drug proprietaries, and Indonesia druggists' sundries
261 KCP Income Fund Perfumes, cosmetics, and other Canada 215 17/05/05 CCL Industries Inc- Metal cans United States toilet preparations North Amer Kemira Oyj Chemicals and chemical Finland 191 06/04/05 Verdugt BV Chemicals and chemical Netherlands preparations, nec preparations, nec Kemira Oyj Chemicals and chemical Finland 444 01/04/05 Finnish Chemicals Chemicals and chemical Finland preparations, nec Oy preparations, nec Kemira Oyj Chemicals and chemical Finland 138 30/01/02 Vinings Industries Industrial organic chemicals, nec United States preparations, nec King Pharmaceuticals Inc Pharmaceutical preparations United States 750 13/06/03 Elan Corp PLC- Pharmaceutical preparations United States 1 750 Primary Care King Pharmaceuticals Inc Pharmaceutical preparations United States 235 08/01/03 Meridian Medical Electromedical and United States 1 1 235 235 Technologies electrotherapeutic apparatus King Pharmaceuticals Inc Pharmaceutical preparations United States 275 31/12/02 Aventis- Pharmaceutical preparations France 1 1 275 275 Intale,Tilade,Suner cid King Pharmaceuticals Inc Pharmaceutical preparations United States 115 29/05/02 Ortho-McNeil Pharmaceutical preparations United States 1 115 Pharmaceutical King Pharmaceuticals Inc Pharmaceutical preparations United States 285 09/08/01 Bristol-Myers Pharmaceutical preparations United States 1 285 Squibb-US Rights King Pharmaceuticals Inc Pharmaceutical preparations United States 3363 31/08/00 Jones Pharmaceutical preparations United States 1 3363 Pharmaceutical Inc
King Pharmaceuticals Inc Pharmaceutical preparations United States 356 25/02/00 Medco Research Pharmaceutical preparations United States 1 356 Inc King Pharmaceuticals Inc Pharmaceutical preparations United States 363 22/12/98 Hoechst Marion Pharmaceutical preparations Germany 1 1 363 363 Roussel-Prods King 8 2 1 5741 637 235 Knoll Pharmaceutical preparations United States 450 05/03/01 Hokuriku Seiyaku Pharmaceutical preparations Japan Pharmaceuticals(Abbott) Co Ltd Koninklijke Numico NV Dry, condensed, and Netherlands 529 24/06/05 Mellin SpA Canned specialties Italy evaporated dairy products Koninklijke Numico NV Dry, condensed, and Netherlands 1747 10/07/00 Rexall Sundown Pharmaceutical preparations United States evaporated dairy products Inc Koninklijke Numico NV Dry, condensed, and Netherlands 2546 11/08/99 General Nutrition Miscellaneous food stores United States evaporated dairy products Cos Inc Kos Pharmaceuticals Inc Pharmaceutical preparations United States 200 05/03/04 Aventis Pharm- Pharmaceutical preparations United States Azmacort Rights Kowa Co Ltd Pharmaceutical preparations Japan 130 13/11/03 Nikken Chemicals Pharmaceutical preparations Japan Co Ltd Kuraray Co Ltd Chemicals and chemical Japan 238 17/01/02 Clariant AG- Plastics materials and synthetic Switzerland preparations, nec PVA/PVB resins Mallinckrodt Inc In vitro and in vivo diagnostic United States 1864 15/09/97 Nellcor Puritan- Electromedical and United States substances Bennett electrotherapeutic apparatus
262 Marion Merrell Dow Inc Pharmaceutical preparations United States 275 05/10/93 Rugby-Darby Medical, dental, and hospital United States Group Cos-Drug equipment & supplies Bus Matrix Laboratories Ltd Pharmaceutical preparations India 203 04/10/05 Docpharma NV Pharmaceutical preparations Belgium Mayne Group Ltd Pharmaceutical preparations Australia 105 26/04/04 aaiPharma- Pharmaceutical preparations United States Injectable Prod Mayne Group Ltd Pharmaceutical preparations Australia 153 29/09/02 Queensland Med Medical laboratories Australia Laboratory Grp Meda AB Pharmaceutical preparations Sweden 909 28/09/05 VIATRIS GmbH & Pharmaceutical preparations Germany Co KG Meda AB Pharmaceutical preparations Sweden 135 20/01/05 Novartis AG-Brand Pharmaceutical preparations Switzerland Rights(2) Medeus Pharma Ltd Pharmaceutical preparations United 120 12/02/04 Elan-Certain Pharmaceutical preparations Ireland-Rep Kingdom European Bus Medicis Pharmaceutical Pharmaceutical preparations United States 160 10/03/03 HA North American In vitro and in vivo diagnostic United States Corp Sales AB substances
Medicis Pharmaceutical Pharmaceutical preparations United States 136 15/11/01 Ascent Pediatrics Pharmaceutical preparations United States Corp Inc MedImmune Inc Biological products, except United States 1740 15/01/02 Aviron Biological products, except United States diagnostic substances diagnostic substances MedImmune Inc Biological products, except United States 393 23/11/99 US Bioscience Inc Pharmaceutical preparations United States diagnostic substances Merck & Co Inc Pharmaceutical preparations United States 461 19/07/01 Rosetta Commercial physical and United States 1 461 Inpharmatics Inc biological research Merck & Co Inc Pharmaceutical preparations United States 185 21/06/00 Provantage Health Management consulting services United States 1 185 Services Merck & Co Inc Pharmaceutical preparations United States 5921 18/11/93 Medco Drugs, drug proprietaries, and United States 1 5921 Containment druggists' sundries Services Inc Merc 3 0 0 6567 0 0 Merck KGaA Pharmaceutical preparations Germany 935 28/07/99 VWR Scientific Professional equipment and United States 1 935 Products Corp supplies, nec Merck KGaA Pharmaceutical preparations Germany 225 02/05/96 Seven Seas Medicinal chemicals and botanical United 1 225 Ltd(Hanson PLC) products Kingdom Merck KGaA Pharmaceutical preparations Germany 1391 17/10/95 Merck AG Pharmaceutical preparations Switzerland 1 1391 MerK 3 0 0 2551 0 0 MGI PHARMA Inc Pharmaceutical preparations United States 203 03/10/05 Guilford Pharmaceutical preparations United States Pharmaceuticals Inc Miles Inc Pharmaceutical preparations United States 1000 03/11/94 Sterling Winthrop- Drugs, drug proprietaries, and United States NA OTC Drug druggists' sundries Miles Inc Pharmaceutical preparations United States 101 18/04/94 ChemDesign Industrial organic chemicals, nec United States Corp(Bayer Corp)
263 Millennium Pharmaceutical preparations United States 2174 12/02/02 COR Therapeutics Pharmaceutical preparations United States Pharmaceuticals Inc Inc Millennium Pharmaceutical preparations United States 557 22/12/99 LeukoSite Inc Pharmaceutical preparations United States Pharmaceuticals Inc Millipore Corp Laboratory analytical United States 151 27/01/97 Tylan General Inc Process control instruments United States instruments Millipore Corp Laboratory analytical United States 125 31/12/96 Amicon Inc(Natl Laboratory analytical instruments United States instruments Med Care Inc) Monsanto Co Pesticides and agricultural United States 26772 31/03/00 Pharmacia & Pharmaceutical preparations United States chemicals, nec Upjohn Inc Monsanto Co Pesticides and agricultural United States 2382 07/12/98 DeKalb Genetics Commercial physical and United States chemicals, nec Corp biological research Monsanto Co Pesticides and agricultural United States 1400 30/10/98 Cargill-International Grain and field beans Mexico chemicals, nec Seed Ope
Monsanto Co Pesticides and agricultural United States 523 16/07/98 Plant Breeding Intl Ornamental floriculture and United chemicals, nec Cambridge nursery products Kingdom Monsanto Co Pesticides and agricultural United States 945 04/09/97 Holden's Ornamental floriculture and United States chemicals, nec Foundation Seeds nursery products Monsanto Co Pesticides and agricultural United States 243 21/05/97 Calgene Ornamental floriculture and United States chemicals, nec Inc(Monsanto Co) nursery products Monsanto Co Pesticides and agricultural United States 240 03/02/97 Asgrow Ornamental floriculture and United States chemicals, nec Agronomics(Semin nursery products is) Monsanto Co Pesticides and agricultural United States 150 21/05/96 Agracetus- Commercial physical and United States chemicals, nec Transgenic Plant biological research Bus Monsanto Co Pesticides and agricultural United States 1075 21/02/95 Kelco Biopolymers Industrial organic chemicals, nec United States chemicals, nec Monsanto Co Pesticides and agricultural United States 400 14/05/93 Chevron Chemical Pesticides and agricultural United States chemicals, nec Co-Ortho chemicals, nec Mylan Laboratories Inc Pharmaceutical preparations United States 188 05/10/98 Penederm Inc Pharmaceutical preparations United States Nabi Biopharmaceuticals Biological products, except United States 101 05/08/03 Braintree Labs Inc- Pharmaceutical preparations United States diagnostic substances PhosLo Natraceutical SA Biological products, except Spain 104 04/05/05 Braes Group Ltd Flavoring extracts and flavoring United diagnostic substances syrups, nec Kingdom NBTY Inc Pharmaceutical preparations United States 115 01/08/05 Solgar Vitamin & Medicinal chemicals and botanical United States Herb Co products NBTY Inc Pharmaceutical preparations United States 250 25/07/03 Rexall Sundown Pharmaceutical preparations United States Inc NBTY Inc Pharmaceutical preparations United States 169 08/08/97 Holland & Miscellaneous food stores United Barrett(Lloyds) Kingdom NeoSan Pharm(AaiPharma Pharmaceutical preparations United States 100 30/08/01 Astrazeneca AB- Pharmaceutical preparations Sweden Inc) Critical Care
264 North American Biologicals Biological products, except United States 160 30/11/95 Univax Biologics Biological products, except United States Inc diagnostic substances Inc diagnostic substances NOVA Chemicals Corp Plastics materials and synthetic Canada 185 31/01/00 Royal Dutch/Shell Plastics materials and synthetic Netherlands resins Group- resins Novartis AG Pharmaceutical preparations Switzerland 660 31/08/05 Bristol-Myers Pharmaceutical preparations United States 1 1 660 660 Squibb Co-US Novartis AG Pharmaceutical preparations Switzerland 933 26/07/05 Eon Labs Inc Pharmaceutical preparations United States 1 1 933 933 Novartis AG Pharmaceutical preparations Switzerland 1504 21/07/05 Eon Labs Inc Pharmaceutical preparations United States 1 1 1504 1504 Novartis AG Pharmaceutical preparations Switzerland 5685 06/06/05 Hexal AG Pharmaceutical preparations Germany 1 1 5685 5685 Novartis AG Pharmaceutical preparations Switzerland 225 01/04/03 Pfizer Inc-Enablex Pharmaceutical preparations United States 1 1 225 225 Brand Novartis AG Pharmaceutical preparations Switzerland 851 18/11/02 Lek(Slovenia) Pharmaceutical preparations Slovenia 1 1 851 851 Novartis AG Pharmaceutical preparations Switzerland 421 24/04/01 Hazal Investment advice France 1 1 1 421 421 421 Finance(Negma) Novartis AG Pharmaceutical preparations Switzerland 1634 21/12/00 SB- Pharmaceutical preparations United 1 1 1634 1634 Famvir,Vectavir/De Kingdom navir Novartis AG Pharmaceutical preparations Switzerland 143 31/08/98 Oriental Chemical Pesticides and agricultural South Korea 1 1 1 143 143 143 Inds-Crop chemicals, nec Novartis AG Pharmaceutical preparations Switzerland 910 03/07/97 Merck-Crop Pesticides and agricultural United States 1 1 1 910 910 910 Protection chemicals, nec Business Novartis Generics(Novartis Pharmaceutical preparations Austria 101 01/01/01 BASF Pharma- Pharmaceutical preparations Germany 1 1 101 101 AG) Euro Generics Bus
Novartis Medical Nutrition Medicinal chemicals and Switzerland 385 17/02/04 Mead Johnson- Food preparations, nec United States 1 1 1 385 385 385 botanical products Adult Nut Bus Novartis Pharma AG Pharmaceutical preparations Switzerland 612 09/05/03 Idenix Pharmaceutical preparations United States 1 1 612 612 Pharmaceuticals Inc Sandoz GmbH Pharmaceutical preparations Austria 565 16/08/04 Sabex Inc Pharmaceutical preparations Canada 1 1 565 565 Nova 14 14 4 14629 14629 1859 Omega Pharma NV Pharmaceutical preparations Belgium 122 03/09/04 Medestea Pharmaceutical preparations Italy International Srl Omega Pharma NV Pharmaceutical preparations Belgium 164 28/06/04 Pfizer-European Pharmaceutical preparations Belgium Brands Omega Pharma NV Pharmaceutical preparations Belgium 118 15/12/00 Chefaro Pharmaceutical preparations Netherlands International(Akzo NV) Omega Pharma NV Pharmaceutical preparations Belgium 139 08/09/00 Fagron Drugs, drug proprietaries, and Netherlands Farmaceuticals(Fa druggists' sundries gron) Omnicare Inc Drug stores and proprietary United States 235 15/08/05 RxCrossroads LLC Health and allied services, nec United States stores
265 Omnicare Inc Drug stores and proprietary United States 269 12/08/05 ExcelleRx Inc Pharmaceutical preparations United States stores Omnicare Inc Drug stores and proprietary United States 2067 28/07/05 NeighborCare Inc Skilled nursing care facilities United States stores Omnicare Inc Drug stores and proprietary United States 402 16/01/03 NCS HealthCare Drug stores and proprietary stores United States stores Inc Omnicare Inc Drug stores and proprietary United States 115 08/01/02 American Drug stores and proprietary stores United States stores Pharmaceutical Svcs Omnicare Inc Drug stores and proprietary United States 255 17/09/98 United Drugs, drug proprietaries, and United States stores Professional Cos druggists' sundries Omnicare Inc Drug stores and proprietary United States 152 29/06/98 IBAH Inc Pharmaceutical preparations United States stores Omnicare Inc Drug stores and proprietary United States 252 16/09/97 American Drugs, drug proprietaries, and United States stores Medserve Corp druggists' sundries Oriental Chemical Inds Co Chemicals and chemical South Korea 208 30/04/00 Korea Steel Steel works, blast furnaces, and South Korea Ltd preparations, nec Chem(Pohang rolling mills Iron) Ortho Biotech Products LP Biological products, except United States 134 09/08/01 Pharmamar Pharmaceutical preparations Spain diagnostic substances Ortho-McNeil Pharm Inc Pharmaceutical preparations United States 245 30/06/05 Peninsula Biological products, except United States Pharmaceuticals diagnostic substances Inc OSI Pharmaceuticals Inc Pharmaceutical preparations United States 721 14/11/05 Eyetech Pharmaceutical preparations United States Pharmaceuticals Inc OSI Pharmaceuticals Inc Pharmaceutical preparations United States 200 21/12/01 Gilead Sciences Pharmaceutical preparations United States Inc-Oncology A PAREXEL International Biological products, except United States 109 01/03/98 PPS Europe Ltd Management consulting services United Corp diagnostic substances Kingdom Patheon Inc Pharmaceutical preparations Canada 442 23/12/04 Mova Pharmaceutical preparations Puerto Rico Pharmaceuticals Corp Perrigo Co Biological products, except United States 922 17/03/05 Agis Pharmaceutical preparations Israel diagnostic substances Industries(1983)Ltd
Pfizer Inc Pharmaceutical preparations United States 1791 14/09/05 Vicuron Biological products, except United States 1 1791 Pharmaceuticals diagnostic substances Inc Pfizer Inc Pharmaceutical preparations United States 527 05/05/05 Angiosyn Inc Biological products, except United States 1 527 diagnostic substances Pfizer Inc Pharmaceutical preparations United States 118 12/11/04 Meridica Ltd Pharmaceutical preparations United 1 1 118 118 Kingdom Pfizer Inc Pharmaceutical preparations United States 372 02/11/04 Slough-Global Commercial physical and United States 1 372 Research Center biological research
266 Pfizer Inc Pharmaceutical preparations United States 620 01/10/04 Aventis SA- Pharmaceutical preparations France 1 1 620 620 Campto Cancer Drug Pfizer Inc Pharmaceutical preparations United States 126 26/03/04 CSL Ltd-Animal Biological products, except Australia 1 1 126 126 Health Business diagnostic substances Pfizer Inc Pharmaceutical preparations United States 1198 11/02/04 Esperion Pharmaceutical preparations United States 1 1198 Therapeutics Inc Pfizer Inc Pharmaceutical preparations United States 60704 15/04/03 Pharmacia Corp Pharmaceutical preparations United States 1 60704 Pfizer Inc Pharmaceutical preparations United States 88771 19/06/00 Warner-Lambert Pharmaceutical preparations United States 1 88771 Co Pfizer Inc Pharmaceutical preparations United States 156 16/03/95 NAMIC USA Corp Surgical and medical instruments United States 1 1 156 156 and apparatus Pfizer Inc Pharmaceutical preparations United States 1450 19/01/95 SmithKline Drugs, drug proprietaries, and United States 1 1450 Beecham Animal druggists' sundries Hlth Pharmacia & Upjohn Inc Pharmaceutical preparations United States 613 31/08/99 SUGEN Inc Commercial physical and United States 1 613 biological research Pharmacia Corp Pharmaceutical preparations United States 200 01/07/02 AT&T Corp- Operators of nonresidential United States 1 1 200 200 Headquarters,Bask buildings ing Upjohn Co Pharmaceutical preparations United States 6802 02/11/95 Pharmacia AB Pharmaceutical preparations Sweden 1 1 6802 6802 Pfiz 14 4 2 163448 7667 356 Pharm Prod Dvlp Inc Commercial physical and United States 481 26/09/96 Applied Bioscience Testing laboratories United States biological research Intl(IMS)
Pharmaceutical Resources Pharmaceutical preparations United States 145 10/06/04 Kali Laboratories Pharmaceutical preparations United States Inc Inc Pharmacopeia Inc Biological products, except United States 127 14/06/98 Molecular Prepackaged Software United States diagnostic substances Simulations Inc Phoenix Int Beteligungs Pharmaceutical preparations Germany 231 15/12/03 Tamro Oyj Drugs, drug proprietaries, and Finland GmbH druggists' sundries Phoenix Int Beteligungs Pharmaceutical preparations Germany 102 14/08/03 Tamro Oyj Drugs, drug proprietaries, and Finland GmbH druggists' sundries PLIVA dd Pharmaceutical preparations Croatia 212 22/06/02 Sobel USA Pharmaceutical preparations United States Inc(Sobel BV) Prestige Brands Perfumes, cosmetics, and other United States 335 10/01/03 Abbott Pharmaceutical preparations United States International toilet preparations Laboratories- Murine Probitas Pharma SA Pharmaceutical preparations Spain 149 25/09/01 SeraCare Inc Specialty outpatient facilities, nec United States
Procter & Gamble Co Soap & other detergents, United States 57227 01/10/05 Gillette Co Cutlery United States except specialty cleaners Procter & Gamble Co Soap & other detergents, United States 208 03/07/04 Laboratorios Vita- Pharmaceutical preparations Spain except specialty cleaners Commercial
267 Procter & Gamble Co Soap & other detergents, United States 2000 30/06/04 Procter & Gamble- Soap & other detergents, except China except specialty cleaners Hutchison Ltd specialty cleaners Procter & Gamble Co Soap & other detergents, United States 1214 30/06/04 Wella AG Perfumes, cosmetics, and other Germany except specialty cleaners toilet preparations Procter & Gamble Co Soap & other detergents, United States 1591 10/09/03 Wella AG Perfumes, cosmetics, and other Germany except specialty cleaners toilet preparations Procter & Gamble Co Soap & other detergents, United States 4530 02/09/03 Wella AG Perfumes, cosmetics, and other Germany except specialty cleaners toilet preparations Procter & Gamble Co Soap & other detergents, United States 4950 16/11/01 Bristol-Myers Perfumes, cosmetics, and other United States except specialty cleaners Squibb-Clairol toilet preparations Procter & Gamble Co Soap & other detergents, United States 259 08/10/99 Recovery Service industry machines, nec United States except specialty cleaners Engineering Inc Procter & Gamble Co Soap & other detergents, United States 2300 01/09/99 IAMs Co Dog, cat, and pet food United States except specialty cleaners Procter & Gamble Co Soap & other detergents, United States 113 17/08/99 Long Chen Paper Paper mills Taiwan except specialty cleaners Co Ltd- Procter & Gamble Co Soap & other detergents, United States 375 15/04/99 Prosan(CMPC,Pro Paper mills Argentina except specialty cleaners cter & Gamble) Procter & Gamble Co Soap & other detergents, United States 170 31/12/97 Loreta y Pena Paper mills Mexico except specialty cleaners Pobre SA de CV Procter & Gamble Co Soap & other detergents, United States 169 26/11/97 Ssangyong Paper Sanitary paper products South Korea except specialty cleaners Co Procter & Gamble Co Soap & other detergents, United States 1976 21/07/97 Tambrands Inc Sanitary paper products United States except specialty cleaners Procter & Gamble Co Soap & other detergents, United States 220 28/06/96 Kimberly-Clark-4 Sanitary paper products United States except specialty cleaners Businesses Procter & Gamble Co Soap & other detergents, United States 150 29/08/94 Giorgio Beverly Perfumes, cosmetics, and other United States except specialty cleaners Hills(Avon) toilet preparations Procyon Biopharma Inc Biological products, except Canada 155 21/04/03 Pharmacor Inc In vitro and in vivo diagnostic Canada diagnostic substances substances Protein Design Labs Inc Biological products, except United States 509 24/03/05 ESP Pharma Inc Pharmaceutical preparations United States diagnostic substances Protein Design Labs Inc Biological products, except United States 108 07/04/03 Eos Biotechnology Pharmaceutical preparations United States diagnostic substances Proteo Inc(Proteo Mkgt Commercial physical and United States 183 25/04/02 Proteo Marketing Pharmaceutical preparations United States Inc) biological research Inc pSiVida Ltd Pharmaceutical preparations Australia 104 13/10/05 Control Delivery Pharmaceutical preparations United States Systems Inc Qiagen NV Biological products, except Netherlands 120 30/06/00 Operon Biological products, except United States diagnostic substances Technologies Inc diagnostic substances QLT Inc Biological products, except Canada 734 19/11/04 Atrix Laboratories Commercial physical and United States diagnostic substances Inc biological research Recordati SpA Pharmaceutical preparations Italy 102 28/06/00 Bouchara SA Pharmaceutical preparations France Rengo Co Ltd Corrugated and solid fiber Japan 638 01/04/99 Settsu Corp Paperboard mills Japan boxes
268 Revco DS Inc Pharmaceutical preparations United States 379 23/12/96 Big B Inc Drug stores and proprietary stores United States
Revco DS Inc Pharmaceutical preparations United States 658 15/07/94 Hook-SupeRx Inc Drug stores and proprietary stores United States
Rexall Sundown Inc Pharmaceutical preparations United States 108 10/01/00 MET-Rx Nutrition Medicinal chemicals and botanical United States Inc products Roche Holding AG Pharmaceutical preparations Switzerland 181 25/07/05 GlycArt Biological products, except Switzerland 1 181 Biotechnology AG diagnostic substances Roche Holding AG Pharmaceutical preparations Switzerland 1254 13/02/04 IGEN International Pharmaceutical preparations United States 1 1 1254 1254 Inc Roche Holding AG Pharmaceutical preparations Switzerland 1189 28/11/03 Disetronic Holding Surgical and medical instruments Switzerland 1 1 1189 1189 AG and apparatus Roche Holding AG Pharmaceutical preparations Switzerland 1230 31/12/00 SmithKline Pharmaceutical preparations United 1 1 1230 1230 Beecham PLC- Kingdom Kytril Roche Holding AG Pharmaceutical preparations Switzerland 4313 16/06/99 Genentech Inc Biological products, except United States 1 1 4313 4313 diagnostic substances Roche Holding AG Pharmaceutical preparations Switzerland 10200 05/03/98 Corange Ltd Pharmaceutical preparations Bermuda 1 1 10200 10200 Roche Holding AG Pharmaceutical preparations Switzerland 1100 31/03/97 Tastemaker Industrial organic chemicals, nec United States 1 1 1100 1100 Roche Holding AG Pharmaceutical preparations Switzerland 5371 03/11/94 Syntex Corp Pharmaceutical preparations United States 1 5371 Roche Holding AG Pharmaceutical preparations Switzerland 141 08/02/93 Fisons PLC- Perfumes, cosmetics, and other Australia 1 1 1 141 141 141 Australian,New toilet preparations Hoffmann-La Roche Inc Pharmaceutical preparations United States 150 31/07/02 Memory Pharmaceutical preparations United States 1 150 Pharmaceuticals Corp- RocH 10 5 3 25129 17138 2430 Roussel-Uclaf SA Pharmaceutical preparations France 140 04/07/95 Dow Chemical Co- Pharmaceutical preparations Brazil Latin American Roussel-Uclaf SA Pharmaceutical preparations France 239 11/02/94 Albert Roussel Pharmaceutical preparations Germany Pharma,1 other Salix Pharmaceuticals Ltd Pharmaceutical preparations United States 182 30/09/05 InKine Biological products, except United States Pharmaceutical Co diagnostic substances
Sanofi-Aventis SA Pharmaceutical preparations France 664 12/07/05 Hoechst AG Manmade organic fibers, except Germany 1 1 1 664 664 664 cellulosic Sanofi-Synthelabo SA Pharmaceutical preparations France 65657 20/08/04 Aventis SA Pharmaceutical preparations France 1 65657 Rhone-Poulenc Rorer Inc Pharmaceutical preparations United States 2559 20/10/95 Fisons PLC Pharmaceutical preparations United 1 1 2559 2559 Kingdom Rhone-Poulenc Rorer Inc Pharmaceutical preparations United States 150 13/02/95 Applied Immune Surgical and medical instruments United States 1 1 150 150 Sciences Inc and apparatus Elf Sanofi SA Pharmaceutical preparations France 1825 03/10/94 Sterling Winthrop- Drugs, drug proprietaries, and United States 1 1 1825 1825 Prescription druggists' sundries Elf Sanofi SA Pharmaceutical preparations France 1003 18/06/93 Yves Saint Laurent Men's and boys' clothing, nec France 1 1 1003 1003 SA Sano 6 71858
269 Saturn Pharmaceuticals Pharmaceutical preparations United States 140 01/07/05 Odyssey Pharm Pharmaceutical preparations United States Inc Inc-Sanctura Schein Pharmaceutical Inc Pharmaceutical preparations United States 229 01/09/95 Marsam Pharmaceutical preparations United States Pharmaceuticals Inc Schering AG Pharmaceutical preparations Germany 380 16/07/02 Immunex Corp- Pharmaceutical preparations United States 1 1 380 380 Leukine Business Schering AG Pharmaceutical preparations Germany 137 03/07/02 Collateral Commercial physical and United States 1 1 137 137 Therapeutics Inc biological research Schering AG Pharmaceutical preparations Germany 314 02/08/96 Leiras(Huhtamaki Biological products, except Finland 1 1 314 314 Oy) diagnostic substances Schering AG Pharmaceutical preparations Germany 336 23/07/96 Jenapharm Medicinal chemicals and botanical Germany 1 1 336 336 GmbH(Gehe AG) products Schering-Plough Corp Pharmaceutical preparations United States 405 01/07/97 Mallinckrodt Prepared animal feeds, except for United States 1 1 405 405 Veterinary Inc dogs and cats Schr 5 4 1 1572 1167 405 Schwarz Pharma AG Pharmaceutical preparations Germany 116 15/08/95 Reed & Pharmaceutical preparations United States Carnrick(Block Drug Co) Schwarz Pharma Kremers- Pharmaceutical preparations United States 178 06/06/95 Central Pharmaceutical preparations United States Urban Pharmaceuticals Schwarz Pharma Kremers- Pharmaceutical preparations United States 116 06/06/95 Reed & Carnrick- Pharmaceutical preparations United States Urban Certain Assets Serologicals Corp Biological products, except United States 202 14/10/04 Upstate Group Inc In vitro and in vivo diagnostic United States diagnostic substances substances Serono International SA Biological products, except Switzerland 162 05/11/02 Genset SA Biological products, except France diagnostic substances diagnostic substances Shanghai Indl Hldg Ltd Offices of holding companies, Hong Kong 120 05/07/00 Active Services Investors, nec Hong Kong nec Group Ltd Shield Diagnostics Group In vitro and in vivo diagnostic United 118 27/05/99 Axis Biochemicals Medicinal chemicals and botanical Norway PLC substances Kingdom AS products Shionogi & Co Ltd Pharmaceutical preparations Japan 120 14/01/93 Eli Lilly & Co- Pharmaceutical preparations United States 1 1 120 120 Capsule Bus Shio 1 1 0 120 120 0 Shire Pharmaceuticals Pharmaceutical preparations United 163 29/12/97 Richwood Drugs, drug proprietaries, and United States 1 1 163 163 Group Kingdom Pharmaceutical Co druggists' sundries Inc Shire Pharmaceuticals Pharmaceutical preparations United 171 24/03/97 Pharmavene Inc Pharmaceutical preparations United States 1 1 171 171 Group Kingdom Shire Pharmaceuticals Grp Pharmaceutical preparations United 1347 28/07/05 Transkaryotic Biological products, except United States 1 1 1347 1347 PLC Kingdom Therapies Inc diagnostic substances Shire Pharmaceuticals Grp Pharmaceutical preparations United 3782 11/05/01 BioChem Pharma Pharmaceutical preparations Canada 1 1 3782 3782 PLC Kingdom Inc
270 Shire Pharmaceuticals Grp Pharmaceutical preparations United 1066 23/12/99 Roberts Pharmaceutical preparations United States 1 1 1066 1066 PLC Kingdom Pharmaceutical Corp Shir 5 5 0 6528 6528 0 Sigma Co Ltd Pharmaceutical preparations Australia 513 02/12/05 Arrow Pharmaceutical preparations Australia Pharmaceuticals Ltd Sigma-Aldrich Corp Chemicals and chemical United States 370 01/03/05 JRH Biosciences Biological products, except United States preparations, nec Inc diagnostic substances Sika AG Chemicals and chemical Switzerland 458 19/12/05 Sarna Kunststoff Plastics materials and synthetic Switzerland preparations, nec Holding AG resins SkyePharma PLC Pharmaceutical preparations United 446 03/05/96 Jago Holding AG Offices of holding companies, nec Switzerland Kingdom Snia SpA Industrial organic chemicals, Italy 116 22/01/03 Centerpulse-Heart Orthopedic, prosthetic, and United States nec Valve Bus surgical supplies Solvay Pharmaceuticals Pharmaceutical preparations Belgium 112 21/07/99 Unimed Pharmaceutical preparations United States 1 112 SA Pharmaceuticals Inc Solv 1 0 0 112 0 0 Sorin Biomedica SpA Pharmaceutical preparations Italy 267 18/05/99 COBE Surgical and medical instruments United States Cardiovascular(CO and apparatus BE Lab) Sosei Co Ltd Biological products, except Japan 185 30/08/05 Arakis Ltd Pharmaceutical preparations United diagnostic substances Kingdom SRF Ltd Synthetic rubber (vulcanizable India 103 28/10/96 Ceat Tyres-Nylon Tire cord and fabrics India elastomers) Tyre Cord Uni STADA Arzneimittel AG Pharmaceutical preparations Germany 108 08/02/05 OAO Nizhpharm Pharmaceutical preparations Russian Fed Suzuken Co Ltd Pharmaceutical preparations Japan 160 30/07/98 Akiyama Inc Drugs, drug proprietaries, and Japan druggists' sundries Takeda Pharmaceutical Co Pharmaceutical preparations Japan 270 01/03/05 Syrrx Inc Biological products, except United States 1 1 270 270 Ltd diagnostic substances Take 1 1 0 270 270 0 Talecris Biotherapeutics Pharmaceutical preparations United States 590 01/04/05 NPS Biological products, except United States Hldg BioTherapeutics diagnostic substances Inc Terumo Corp Laboratory analytical Japan 170 18/11/02 Vascutek Orthopedic, prosthetic, and United instruments Ltd(Centerpulse surgical supplies Kingdom AG) Terumo Corp Laboratory analytical Japan 110 30/06/99 3M-Cardiovascular Surgical and medical instruments United States instruments Business and apparatus Teva Pharm Inds Ltd Pharmaceutical preparations Israel 3165 22/01/04 SICOR Inc Pharmaceutical preparations United States 1 3165 Teva Pharm Inds Ltd Pharmaceutical preparations Israel 285 05/04/00 Novopharm Pharmaceutical preparations Canada 1 285 Ltd(Dan Family Hold)
271 Teva Pharm Inds Ltd Pharmaceutical preparations Israel 350 31/05/96 Biocraft Pharmaceutical preparations United States 1 350 Laboratories Inc Teva Pharmaceutical USA Pharmaceutical preparations United States 187 20/09/99 Copley Pharmaceutical preparations United States 1 187 Inc Pharmaceutical Inc
Teva 4 0 0 3988 0 0 UCB SA Pharmaceutical preparations Belgium 2473 06/07/04 Celltech Group Commercial physical and United 1 1 2473 2473 PLC biological research Kingdom UCB SA Pharmaceutical preparations Belgium 500 31/01/03 Solutia Inc- Chemicals and chemical United States 1 1 500 500 Specialty Chem preparations, nec Bus UCB 2 2 0 2973 2973 0
Valeant Pharm Intl Inc Pharmaceutical preparations United States 324 01/03/05 Xcel Pharmaceutical preparations United States Pharmaceuticals Inc Versicor Inc Biological products, except United States 153 03/03/03 Biosearch Italia Biological products, except Italy diagnostic substances SpA diagnostic substances Vertex Pharmaceuticals Inc Pharmaceutical preparations United States 556 18/07/01 Aurora Biosciences Laboratory analytical instruments United States Corp VI Technologies Inc Biological products, except United States 152 11/03/05 Panacos Biological products, except United States diagnostic substances Pharmaceuticals diagnostic substances Inc VIMRx Pharmaceuticals Pharmaceutical preparations United States 120 18/12/97 Baxter Healthcare Medical laboratories United States Inc Corp ViroPharma Inc Pharmaceutical preparations United States 116 10/11/04 Eli Lilly-Vanconcin Pharmaceutical preparations United States Rights Warner Chilcott PLC Pharmaceutical preparations Ireland-Rep 180 16/02/00 Bristol-Myers- Pharmaceutical preparations United States Women's Prods(3) Warner-Lambert Co Pharmaceutical preparations United States 2132 17/05/99 Agouron Pharmaceutical preparations United States Pharmaceuticals Inc Warner-Lambert Co Pharmaceutical preparations United States 1050 01/07/96 Warner Wellcome Drugs, drug proprietaries, and United Consumer Hlth druggists' sundries Kingdom Warner-Lambert Co Pharmaceutical preparations United States 142 22/03/93 Wilkinson Sword Cutlery United Group Ltd Kingdom Watson Pharmaceuticals Pharmaceutical preparations United States 178 12/02/03 Novatis AG- Pharmaceutical preparations United States 1 178 Inc Fiorinal Brands Watson Pharmaceuticals Pharmaceutical preparations United States 184 17/11/00 Makoff R&D Pharmaceutical preparations United States 1 184 Inc Laboratories Inc Watson Pharmaceuticals Pharmaceutical preparations United States 899 28/08/00 Schein Pharmaceutical preparations United States 1 899 Inc Pharmaceutical Inc
Watson Pharmaceuticals Pharmaceutical preparations United States 297 18/01/99 TheraTech Inc Pharmaceutical preparations United States 1 297 Inc
272 Watson Pharmaceuticals Pharmaceutical preparations United States 131 28/02/97 Oclassen Pharmaceutical preparations United States 1 131 Inc Pharmaceuticals Inc Watson Pharmaceuticals Pharmaceutical preparations United States 571 17/07/95 Circa Pharmaceutical preparations United States 1 571 Inc Pharmaceuticals Inc Wats 6 0 0 2259 0 0 Welfide Corp Pharmaceutical preparations Japan 1207 01/10/01 Mitsubishi-Tokyo Pharmaceutical preparations Japan Pharmaceutica Whittaker Corp Pharmaceutical preparations United States 118 10/04/96 Xyplex Computer terminals United States Inc(Raytheon Co) Xanodyne Pharmaceuticals Pharmaceutical preparations United States 209 25/07/05 aaiPharma Inc- Biological products, except United States Inc Pharm Division diagnostic substances Yamanouchi Pharmaceutical preparations Japan 7223 01/04/05 Fujisawa Pharmaceutical preparations Japan Pharmaceutical Co Pharmaceutical Co Ltd Yoshitomi Pharmaceutical Pharmaceutical preparations Japan 1010 01/04/98 Green Cross Corp Biological products, except Japan Inds diagnostic substances Zentiva NV Pharmaceutical preparations Czech 102 12/10/05 Venoma Holdings Investors, nec Romania Republic Ltd
273